Release: v4.22.1

* Fix tokenizer load from one file

* Add a test

* Style

Co-authored-by: Lysandre <lysandre.debut@reseau.eseo.fr>

.gitattributes

@@ -1,3 +1,4 @@
 *.py	eol=lf
 *.rst	eol=lf
-*.md	eol=lf
+*.md	eol=lf
+*.mdx   eol=lf

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -1,5 +1,5 @@
 name: "\U0001F41B Bug Report"
-description: Submit a bug report to help us import transformers
+description: Submit a bug report to help us improve transformers
 labels: [ "bug" ]
 body:
   - type: textarea
@@ -7,7 +7,6 @@ body:
     attributes:
       label: System Info
       description: Please share your system info with us. You can run the command `transformers-cli env` and copy-paste its output below.
-      render: shell
       placeholder: transformers version, platform, python version, ...
     validations:
       required: true
@@ -118,4 +117,3 @@ body:
     attributes:
       label: Expected behavior
       description: "A clear and concise description of what you would expect to happen."
-      render: shell

.github/ISSUE_TEMPLATE/config.yml

@@ -1,9 +1,12 @@
 blank_issues_enabled: true
 version: 2.1
 contact_links:
+  - name: Model checkpoints on the Hugging Face Hub
+    url: https://huggingface.co/models
+    about: Open a Pull request / Discussion related to a specific model checkpoint directly on the Hugging Face Hub
   - name: Website Related
     url: https://github.com/huggingface/hub-docs/issues
     about: Feature requests and bug reports related to the website
   - name: Forum
     url: https://discuss.huggingface.co/
-    about: General usage questions and community discussions
+    about: General usage questions and community discussions

.github/conda/meta.yaml

@@ -25,7 +25,7 @@ requirements:
     - sacremoses
     - regex !=2019.12.17
     - protobuf
-    - tokenizers >=0.10.1,<0.11.0
+    - tokenizers >=0.11.1,!=0.11.3,<0.13
     - pyyaml >=5.1
   run:
     - python
@@ -40,7 +40,7 @@ requirements:
     - sacremoses
     - regex !=2019.12.17
     - protobuf
-    - tokenizers >=0.10.1,<0.11.0
+    - tokenizers >=0.11.1,!=0.11.3,<0.13
     - pyyaml >=5.1
 
 test:

.github/workflows/add-model-like.yml

@@ -27,7 +27,7 @@ jobs:
         id: cache
         with:
           path: ~/venv/
-          key: v3-tests_model_like-${{ hashFiles('setup.py') }}
+          key: v4-tests_model_like-${{ hashFiles('setup.py') }}
 
       - name: Create virtual environment on cache miss
         if: steps.cache.outputs.cache-hit != 'true'
@@ -41,10 +41,12 @@ jobs:
         run: |
           . ~/venv/bin/activate
           python setup.py develop
-          transformer_loc=$(pip show transformers | grep "Location: " | cut -c11-)
-          transformer_repo_loc=$(pwd .)
-          if [ "$transformer_loc" != "$transformer_repo_loc/src" ]; then
-              echo "transformers is from $transformer_loc but it shoud be from $transformer_repo_loc/src."
+          transformers_install=$(pip list -e | grep transformers)
+          transformers_install_array=($transformers_install)
+          transformers_loc=${transformers_install_array[-1]}
+          transformers_repo_loc=$(pwd .)
+          if [ "$transformers_loc" != "$transformers_repo_loc" ]; then
+              echo "transformers is from $transformers_loc but it shoud be from $transformers_repo_loc/src."
               echo "A fix is required. Stop testing."
               exit 1
           fi

.github/workflows/build-docker-images.yml

@@ -5,6 +5,7 @@ on:
     branches:
       - docker-image*
   repository_dispatch:
+  workflow_call:
   schedule:
     - cron: "0 1 * * *"
 
@@ -39,9 +40,35 @@ jobs:
           push: true
           tags: huggingface/transformers-all-latest-gpu
 
+  latest-with-torch-nightly-docker:
+    name: "Nightly PyTorch + Stable TensorFlow"
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-all-latest-gpu
+          build-args: |
+            REF=main
+            PYTORCH=pre
+          push: true
+          tags: huggingface/transformers-all-latest-torch-nightly-gpu
+
   latest-torch-deepspeed-docker:
     name: "Latest PyTorch + DeepSpeed"
-    needs: latest-docker
     runs-on: ubuntu-latest
     steps:
       -
@@ -66,6 +93,32 @@ jobs:
           push: true
           tags: huggingface/transformers-pytorch-deepspeed-latest-gpu
 
+  nightly-torch-deepspeed-docker:
+    name: "Nightly PyTorch + DeepSpeed"
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-pytorch-deepspeed-nightly-gpu
+          build-args: |
+            REF=main
+          push: true
+          tags: huggingface/transformers-pytorch-deepspeed-nightly-gpu
+
   doc-builder:
     name: "Doc builder"
     runs-on: ubuntu-latest
@@ -93,7 +146,6 @@ jobs:
   latest-pytorch:
     name: "Latest PyTorch [dev]"
     runs-on: ubuntu-latest
-    needs: latest-torch-deepspeed-docker
     steps:
       -
         name: Set up Docker Buildx
@@ -118,7 +170,6 @@ jobs:
           tags: huggingface/transformers-pytorch-gpu
 
   latest-tensorflow:
-    needs: latest-pytorch
     name: "Latest TensorFlow [dev]"
     runs-on: ubuntu-latest
     steps:

.github/workflows/build-past-ci-docker-images.yml

@@ -0,0 +1,108 @@
+name: Build docker images (Past CI)
+
+on:
+  push:
+    branches:
+      - past-ci-docker-image*
+
+concurrency:
+  group: docker-images-builds
+  cancel-in-progress: false
+
+jobs:
+  past-pytorch-docker:
+    name: "Past PyTorch Docker"
+    strategy:
+      fail-fast: false
+      matrix:
+        version: ["1.11", "1.10", "1.9", "1.8", "1.7", "1.6", "1.5", "1.4"]
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-past-gpu
+          build-args: |
+            REF=main
+            FRAMEWORK=pytorch
+            VERSION=${{ matrix.version }}
+          push: true
+          tags: huggingface/transformers-pytorch-past-${{ matrix.version }}-gpu
+
+  past-tensorflow-docker:
+    name: "Past TensorFlow Docker"
+    strategy:
+      fail-fast: false
+      matrix:
+        version: ["2.8", "2.7", "2.6", "2.5"]
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-past-gpu
+          build-args: |
+            REF=main
+            FRAMEWORK=tensorflow
+            VERSION=${{ matrix.version }}
+          push: true
+          tags: huggingface/transformers-tensorflow-past-${{ matrix.version }}-gpu
+
+  past-tensorflow-docker-2-4:
+    name: "Past TensorFlow Docker"
+    strategy:
+      fail-fast: false
+      matrix:
+        version: ["2.4"]
+    runs-on: ubuntu-latest
+    steps:
+      -
+        name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v1
+      -
+        name: Check out code
+        uses: actions/checkout@v2
+      -
+        name: Login to DockerHub
+        uses: docker/login-action@v1
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_PASSWORD }}
+      -
+        name: Build and push
+        uses: docker/build-push-action@v2
+        with:
+          context: ./docker/transformers-past-gpu
+          build-args: |
+            REF=main
+            BASE_DOCKER_IMAGE=nvidia/cuda:11.0.3-cudnn8-devel-ubuntu20.04
+            FRAMEWORK=tensorflow
+            VERSION=${{ matrix.version }}
+          push: true
+          tags: huggingface/transformers-tensorflow-past-${{ matrix.version }}-gpu

.github/workflows/build_documentation.yml

@@ -15,6 +15,6 @@ jobs:
       commit_sha: ${{ github.sha }}
       package: transformers
       notebook_folder: transformers_doc
-      languages: en es
+      languages: de en es it pt
     secrets:
       token: ${{ secrets.HUGGINGFACE_PUSH }}

.github/workflows/build_pr_documentation.yml

@@ -14,4 +14,4 @@ jobs:
       commit_sha: ${{ github.event.pull_request.head.sha }}
       pr_number: ${{ github.event.number }}
       package: transformers
-      languages: en es
+      languages: de en es it pt

.github/workflows/doctests.yml

@@ -32,9 +32,7 @@ jobs:
 
       - name: GPU visibility
         run: |
-          utils/print_env_pt.py
-          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
+          python3 utils/print_env.py
 
       - name: Prepare files for doctests
         run: |

.github/workflows/model-templates.yml

@@ -21,7 +21,7 @@ jobs:
         id: cache
         with:
           path: ~/venv/
-          key: v3-tests_templates-${{ hashFiles('setup.py') }}
+          key: v4-tests_templates-${{ hashFiles('setup.py') }}
 
       - name: Create virtual environment on cache miss
         if: steps.cache.outputs.cache-hit != 'true'

.github/workflows/self-nightly-scheduled.yml

@@ -1,250 +1,264 @@
-name: Self-hosted runner; Nightly (scheduled)
+name: Self-hosted runner (nightly)
+
+# Note that each job's dependencies go into a corresponding docker file.
+#
+# For example for `run_all_tests_torch_cuda_extensions_gpu` the docker image is
+# `huggingface/transformers-pytorch-deepspeed-latest-gpu`, which can be found at
+# `docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile`
 
 on:
-    push:
-        branches:
-            - nightly_ci*
-    repository_dispatch:
-    schedule:
-        - cron: "0 0 */3 * *"
+  repository_dispatch:
+# Disable temporarily until the test suite can be run under 12 hours.
+#  schedule:
+#    - cron: "0 16 * * *"
 
 env:
-    HF_HOME: /mnt/cache
-    TRANSFORMERS_IS_CI: yes
-    RUN_SLOW: yes
-    OMP_NUM_THREADS: 16
-    MKL_NUM_THREADS: 16
-    PYTEST_TIMEOUT: 600
-    SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  HF_HOME: /mnt/cache
+  TRANSFORMERS_IS_CI: yes
+  OMP_NUM_THREADS: 8
+  MKL_NUM_THREADS: 8
+  RUN_SLOW: yes
+  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  TF_FORCE_GPU_ALLOW_GROWTH: true
+  RUN_PT_TF_CROSS_TESTS: 1
 
 jobs:
-    run_all_tests_torch_gpu:
-        runs-on: [self-hosted, docker-gpu, single-gpu]
-        container:
-            image: pytorch/pytorch:1.10.0-cuda11.3-cudnn8-runtime
-            options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-        steps:
-            - name: Launcher docker
-              uses: actions/checkout@v2
+  run_check_runners:
+    name: Check Runners
+    strategy:
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    container:
+      image: huggingface/transformers-all-latest-torch-nightly-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
 
-            - name: NVIDIA-SMI
-              run: |
-                  nvidia-smi
+  setup:
+    name: Setup
+    needs: run_check_runners
+    strategy:
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    container:
+      image: huggingface/transformers-all-latest-torch-nightly-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    outputs:
+      matrix: ${{ steps.set-matrix.outputs.matrix }}
+    steps:
+      - name: Update clone
+        working-directory: /transformers
+        run: |
+          git fetch && git checkout ${{ github.sha }}
 
-            - name: Install dependencies
-              run: |
-                  apt -y update && apt install -y libsndfile1-dev git espeak-ng
-                  pip install --upgrade pip
-                  pip install .[integrations,sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-                  pip install https://github.com/kpu/kenlm/archive/master.zip
-                  pip install --pre torch torchvision torchaudio -f https://download.pytorch.org/whl/nightly/cu113/torch_nightly.html -U
+      - name: Cleanup
+        working-directory: /transformers
+        run: |
+          rm -rf tests/__pycache__
+          rm -rf tests/models/__pycache__
+          rm -rf reports
 
-            - name: Are GPUs recognized by our DL frameworks
-              run: |
-                utils/print_env_pt.py
+      - id: set-matrix
+        name: Identify models to test
+        working-directory: /transformers/tests
+        run: |
+          echo "::set-output name=matrix::$(python3 -c 'import os; tests = os.getcwd(); model_tests = os.listdir(os.path.join(tests, "models")); d1 = sorted(list(filter(os.path.isdir, os.listdir(tests)))); d2 = sorted(list(filter(os.path.isdir, [f"models/{x}" for x in model_tests]))); d1.remove("models"); d = d2 + d1; print(d)')"
 
-            - name: Run all tests on GPU
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_gpu tests
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_gpu/failures_short.txt
+  run_tests_single_gpu:
+    name: Model tests
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [single-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    container:
+      image: huggingface/transformers-all-latest-torch-nightly-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [run_check_runners, setup]
+    steps:
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
 
-            - name: Run examples tests on GPU
-              if: ${{ always() }}
-              env:
-                  OMP_NUM_THREADS: 16
-                  MKL_NUM_THREADS: 16
-                  RUN_SLOW: yes
-                  HF_HOME: /mnt/cache
-                  TRANSFORMERS_IS_CI: yes
-              run: |
-                  pip install -r examples/pytorch/_tests_requirements.txt
-                  python -m pytest -n 1 -v --dist=loadfile --make-reports=examples_torch_gpu examples
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/examples_torch_gpu/failures_short.txt
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
 
-            - name: Run all pipeline tests on GPU
-              if: ${{ always() }}
-              env:
-                  RUN_PIPELINE_TESTS: yes
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_torch_pipeline_gpu tests
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_pipeline_gpu/failures_short.txt
+      - name: Run all tests on GPU
+        working-directory: /transformers
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
 
-            - name: Test suite reports artifacts
-              if: ${{ always() }}
-              uses: actions/upload-artifact@v2
-              with:
-                  name: run_all_tests_torch_gpu_test_reports
-                  path: reports
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
-    run_all_tests_torch_multi_gpu:
-        runs-on: [self-hosted, docker-gpu, multi-gpu]
-        container:
-            image: pytorch/pytorch:1.10.0-cuda11.3-cudnn8-runtime
-            options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-        steps:
-            - name: Launcher docker
-              uses: actions/checkout@v2
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
-            - name: NVIDIA-SMI
-              continue-on-error: true
-              run: |
-                  nvidia-smi
+  run_tests_multi_gpu:
+    name: Model tests
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    container:
+      image: huggingface/transformers-all-latest-torch-nightly-gpu
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [run_check_runners, setup]
+    steps:
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
 
-            - name: Install dependencies
-              run: |
-                  apt -y update && apt install -y libsndfile1-dev git espeak-ng
-                  pip install --upgrade pip
-                  pip install .[integrations,sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-                  pip install https://github.com/kpu/kenlm/archive/master.zip
-                  pip install --pre torch torchvision torchaudio -f https://download.pytorch.org/whl/nightly/cu113/torch_nightly.html -U
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
-            - name: Are GPUs recognized by our DL frameworks
-              run: |
-                utils/print_env_pt.py
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
 
-            - name: Run all tests on GPU
-              env:
-                  MKL_SERVICE_FORCE_INTEL: 1
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_multi_gpu tests
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_multi_gpu/failures_short.txt
+      - name: Run all tests on GPU
+        working-directory: /transformers
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
 
-            - name: Run all pipeline tests on GPU
-              if: ${{ always() }}
-              env:
-                  RUN_PIPELINE_TESTS: yes
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=tests_torch_pipeline_multi_gpu tests
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_pipeline_multi_gpu/failures_short.txt
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
-            - name: Test suite reports artifacts
-              if: ${{ always() }}
-              uses: actions/upload-artifact@v2
-              with:
-                  name: run_all_tests_torch_multi_gpu_test_reports
-                  path: reports
+  run_all_tests_torch_cuda_extensions_gpu:
+    name: Torch CUDA extension tests
+    strategy:
+      fail-fast: false
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    needs: [run_check_runners, setup]
+    container:
+      image: huggingface/transformers-pytorch-deepspeed-nightly-gpu
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      - name: Update clone
+        working-directory: /workspace/transformers
+        run: git fetch && git checkout ${{ github.sha }}
 
-    run_all_tests_torch_cuda_extensions_gpu:
-        runs-on: [self-hosted, docker-gpu, single-gpu]
-        container:
-            image: nvcr.io/nvidia/pytorch:21.03-py3
-            options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-        steps:
-            - name: Launcher docker
-              uses: actions/checkout@v2
+      - name: Remove cached torch extensions
+        run: rm -rf /github/home/.cache/torch_extensions/
 
-            - name: NVIDIA-SMI
-              run: |
-                  nvidia-smi
+      # To avoid unknown test failures
+      - name: Pre build DeepSpeed *again*
+        working-directory: /workspace
+        run: |
+          python3 -m pip uninstall -y deepspeed
+          rm -rf DeepSpeed
+          git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build
+          DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
 
-            - name: Install dependencies
-              run: |
-                  apt -y update && apt install -y libaio-dev libsndfile1-dev git espeak-ng
-                  pip install --upgrade pip
-                  pip install --pre torch torchvision torchaudio -f https://download.pytorch.org/whl/nightly/cu113/torch_nightly.html -U
-                  pip install .[deepspeed-testing]
-                  pip install https://github.com/kpu/kenlm/archive/master.zip
-                  pip install git+https://github.com/microsoft/DeepSpeed
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
 
-            - name: Are GPUs recognized by our DL frameworks
-              run: |
-                utils/print_env_pt.py
+      - name: Environment
+        working-directory: /workspace/transformers
+        run: |
+          python utils/print_env.py
 
-            - name: Run all tests on GPU
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
+      - name: Run all tests on GPU
+        working-directory: /workspace/transformers
+        run: |
+          python -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
 
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_cuda_extensions_gpu/failures_short.txt
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
 
-            - name: Test suite reports artifacts
-              if: ${{ always() }}
-              uses: actions/upload-artifact@v2
-              with:
-                  name: run_tests_torch_cuda_extensions_gpu_test_reports
-                  path: reports
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
+          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
 
-    run_all_tests_torch_cuda_extensions_multi_gpu:
-        runs-on: [self-hosted, docker-gpu, multi-gpu]
-        container:
-            image: nvcr.io/nvidia/pytorch:21.03-py3
-            options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-        steps:
-            - name: Launcher docker
-              uses: actions/checkout@v2
+  send_results:
+    name: Send results to webhook
+    runs-on: ubuntu-latest
+    if: always()
+    needs: [run_check_runners, setup, run_tests_single_gpu, run_tests_multi_gpu, run_all_tests_torch_cuda_extensions_gpu]
+    steps:
+      - name: Preliminary job status
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          echo "Runner status: ${{ needs.run_check_runners.result }}"
+          echo "Setup status: ${{ needs.setup.result }}"
 
-            - name: NVIDIA-SMI
-              continue-on-error: true
-              run: |
-                  nvidia-smi
-
-            - name: Install dependencies
-              run: |
-                  apt -y update && apt install -y libaio-dev libsndfile1-dev git espeak-ng
-                  pip install --upgrade pip
-                  pip install --pre torch torchvision torchaudio -f https://download.pytorch.org/whl/nightly/cu113/torch_nightly.html -U
-                  rm -rf ~/.cache/torch_extensions/ # shared between conflicting builds
-                  pip install .[testing,fairscale]
-                  pip install https://github.com/kpu/kenlm/archive/master.zip
-                  pip install git+https://github.com/microsoft/DeepSpeed # testing bleeding edge
-
-            - name: Are GPUs recognized by our DL frameworks
-              run: |
-                utils/print_env_pt.py
-
-            - name: Run all tests on GPU
-              run: |
-                  python -m pytest -n 1 -v --dist=loadfile --make-reports=tests_torch_cuda_extensions_multi_gpu tests/deepspeed tests/extended
-
-            - name: Failure short reports
-              if: ${{ always() }}
-              run: cat reports/tests_torch_cuda_extensions_multi_gpu/failures_short.txt
-
-            - name: Test suite reports artifacts
-              if: ${{ always() }}
-              uses: actions/upload-artifact@v2
-              with:
-                  name: run_tests_torch_cuda_extensions_multi_gpu_test_reports
-                  path: reports
-
-    send_results:
-        name: Send results to webhook
-        runs-on: ubuntu-latest
-        if: always()
-        needs: [
-                run_all_tests_torch_gpu,
-                run_all_tests_torch_multi_gpu,
-                run_all_tests_torch_cuda_extensions_gpu,
-                run_all_tests_torch_cuda_extensions_multi_gpu
-        ]
-        steps:
-            - uses: actions/checkout@v2
-
-            - uses: actions/download-artifact@v2
-
-            - name: Send message to Slack
-              env:
-                  CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
-                  CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
-                  CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
-                  CI_SLACK_CHANNEL_ID_PAST_FUTURE: ${{ secrets.CI_SLACK_CHANNEL_ID_PAST_FUTURE }}
-
-              run: |
-                  pip install slack_sdk
-                  python utils/notification_service.py scheduled nightly-torch
+      - uses: actions/checkout@v2
+      - uses: actions/download-artifact@v2
+      - name: Send message to Slack
+        env:
+          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
+          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
+          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
+          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
+          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_PAST_FUTURE }}
+          CI_EVENT: nightly-build
+          SETUP_STATUS: ${{ needs.setup.result }}
+          RUNNER_STATUS: ${{ needs.run_check_runners.result }}
+        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
+        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
+        run: |
+          pip install slack_sdk
+          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"

.github/workflows/self-past-caller.yml

@@ -0,0 +1,136 @@
+name: Self-hosted runner (past-ci-caller)
+
+on:
+  push:
+    branches:
+      - run-past-ci*
+
+jobs:
+  run_past_ci_pytorch_1-11:
+    name: PyTorch 1.11
+    if: always()
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.11"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-10:
+    name: PyTorch 1.10
+    if: always()
+    needs: [run_past_ci_pytorch_1-11]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.10"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-9:
+    name: PyTorch 1.9
+    if: always()
+    needs: [run_past_ci_pytorch_1-10]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.9"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-8:
+    name: PyTorch 1.8
+    if: always()
+    needs: [run_past_ci_pytorch_1-9]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.8"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-7:
+    name: PyTorch 1.7
+    if: always()
+    needs: [run_past_ci_pytorch_1-8]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.7"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-6:
+    name: PyTorch 1.6
+    if: always()
+    needs: [run_past_ci_pytorch_1-7]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.6"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-5:
+    name: PyTorch 1.5
+    if: always()
+    needs: [run_past_ci_pytorch_1-6]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.5"
+    secrets: inherit
+
+  run_past_ci_pytorch_1-4:
+    name: PyTorch 1.4
+    if: always()
+    needs: [run_past_ci_pytorch_1-5]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: pytorch
+      version: "1.4"
+    secrets: inherit
+
+  run_past_ci_tensorflow_2-8:
+    name: TensorFlow 2.8
+    if: always()
+    needs: [run_past_ci_pytorch_1-4]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: tensorflow
+      version: "2.8"
+    secrets: inherit
+
+  run_past_ci_tensorflow_2-7:
+    name: TensorFlow 2.7
+    if: always()
+    needs: [run_past_ci_tensorflow_2-8]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: tensorflow
+      version: "2.7"
+    secrets: inherit
+
+  run_past_ci_tensorflow_2-6:
+    name: TensorFlow 2.6
+    if: always()
+    needs: [run_past_ci_tensorflow_2-7]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: tensorflow
+      version: "2.6"
+    secrets: inherit
+
+  run_past_ci_tensorflow_2-5:
+    name: TensorFlow 2.5
+    if: always()
+    needs: [run_past_ci_tensorflow_2-6]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: tensorflow
+      version: "2.5"
+    secrets: inherit
+
+  run_past_ci_tensorflow_2-4:
+    name: TensorFlow 2.4
+    if: always()
+    needs: [run_past_ci_tensorflow_2-5]
+    uses: ./.github/workflows/self-past.yml
+    with:
+      framework: tensorflow
+      version: "2.4"
+    secrets: inherit

.github/workflows/self-past.yml

@@ -0,0 +1,216 @@
+name: Self-hosted runner (past)
+
+# Note that each job's dependencies go into a corresponding docker file.
+#
+# For example for `run_all_tests_torch_cuda_extensions_gpu` the docker image is
+# `huggingface/transformers-pytorch-deepspeed-latest-gpu`, which can be found at
+# `docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile`
+
+on:
+  workflow_call:
+    inputs:
+      framework:
+        required: true
+        type: string
+      version:
+        required: true
+        type: string
+
+env:
+  HF_HOME: /mnt/cache
+  TRANSFORMERS_IS_CI: yes
+  OMP_NUM_THREADS: 8
+  MKL_NUM_THREADS: 8
+  RUN_SLOW: yes
+  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
+  TF_FORCE_GPU_ALLOW_GROWTH: true
+  RUN_PT_TF_CROSS_TESTS: 1
+
+jobs:
+  setup:
+    name: Setup
+    runs-on: ubuntu-latest
+    outputs:
+      matrix: ${{ steps.set-matrix.outputs.matrix }}
+    steps:
+      - name: Checkout transformers
+        uses: actions/checkout@v2
+        with:
+          fetch-depth: 2
+
+      - name: Cleanup
+        run: |
+          rm -rf tests/__pycache__
+          rm -rf tests/models/__pycache__
+          rm -rf reports
+
+      - id: set-matrix
+        name: Identify models to test
+        run: |
+          cd tests
+          echo "::set-output name=matrix::$(python3 -c 'import os; tests = os.getcwd(); model_tests = os.listdir(os.path.join(tests, "models")); d1 = sorted(list(filter(os.path.isdir, os.listdir(tests)))); d2 = sorted(list(filter(os.path.isdir, [f"models/{x}" for x in model_tests]))); d1.remove("models"); d = d2 + d1; print(d)')"
+
+  run_check_runners:
+    name: Check Runners
+    needs: setup
+    strategy:
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker-past-ci') }}
+    container:
+      image: huggingface/transformers-${{ inputs.framework }}-past-${{ inputs.version }}-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+  run_tests_single_gpu:
+    name: Model tests
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [single-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker-past-ci') }}
+    container:
+      image: huggingface/transformers-${{ inputs.framework }}-past-${{ inputs.version }}-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [setup, run_check_runners]
+    steps:
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
+
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
+
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
+      - name: Run all tests on GPU
+        working-directory: /transformers
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
+
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
+
+  run_tests_multi_gpu:
+    name: Model tests
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker-past-ci') }}
+    container:
+      image: huggingface/transformers-${{ inputs.framework }}-past-${{ inputs.version }}-gpu
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [setup, run_check_runners]
+    steps:
+      - name: Update clone
+        working-directory: /transformers
+        run: git fetch && git checkout ${{ github.sha }}
+
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
+
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
+      - name: Run all tests on GPU
+        working-directory: /transformers
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
+
+      - name: Failure short reports
+        if: ${{ failure() }}
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
+
+      - name: Test suite reports artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
+
+  send_results:
+    name: Send results to webhook
+    runs-on: ubuntu-latest
+    if: always()
+    needs: [setup, run_check_runners, run_tests_single_gpu, run_tests_multi_gpu]
+    steps:
+      - name: Preliminary job status
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          echo "Runner status: ${{ needs.run_check_runners.result }}"
+          echo "Setup status: ${{ needs.setup.result }}"
+
+      - uses: actions/checkout@v2
+      - uses: actions/download-artifact@v2
+
+      # Create a directory to store test failure tables in the next step
+      - name: Create directory
+        run: mkdir test_failure_tables
+
+      - name: Send message to Slack
+        env:
+          CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
+          CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
+          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
+          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
+          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_PAST_FUTURE }}
+          CI_EVENT: Past CI - ${{ inputs.framework }}-${{ inputs.version }}
+          SETUP_STATUS: ${{ needs.setup.result }}
+          RUNNER_STATUS: ${{ needs.run_check_runners.result }}
+        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
+        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
+        run: |
+          pip install slack_sdk
+          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"
+
+      # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack.
+      - name: Failure table artifacts
+        if: ${{ always() }}
+        uses: actions/upload-artifact@v2
+        with:
+          name: test_failure_tables_${{ inputs.framework }}-${{ inputs.version }}
+          path: test_failure_tables

.github/workflows/self-push-caller.yml

@@ -0,0 +1,52 @@
+# Used to trigger self-push CI
+name: Self-hosted runner (push-caller)
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - "src/**"
+      - "tests/**"
+      - ".github/**"
+      - "templates/**"
+      - "utils/**"
+
+jobs:
+  check-for-setup:
+      runs-on: ubuntu-latest
+      name: Check if setup was changed
+      outputs:
+        changed: ${{ steps.was_changed.outputs.changed }}
+      steps:
+        - uses: actions/checkout@v3
+          with: 
+            fetch-depth: "2"
+        
+        - name: Get changed files
+          id: changed-files
+          uses: tj-actions/changed-files@v22.2
+        
+        - name: Was setup changed 
+          id: was_changed
+          run: |
+            for file in ${{ steps.changed-files.outputs.all_changed_files }}; do
+              if [ `basename "${file}"` = "setup.py" ]; then
+                echo ::set-output name=changed::"1"
+              fi
+            done
+
+  build-docker-containers:
+    needs: check-for-setup
+    if: (github.event_name == 'push') && (needs.check-for-setup.outputs.changed == '1')
+    uses: ./.github/workflows/build-docker-images.yml
+    secrets: inherit
+
+  run_push_ci:
+    name: Trigger Push CI
+    runs-on: ubuntu-latest
+    if: ${{ always() }}
+    needs: build-docker-containers
+    steps:
+      - name: Trigger push CI via workflow_run
+        run: echo "Trigger push CI via workflow_run"

.github/workflows/self-push.yml

@@ -1,9 +1,12 @@
 name: Self-hosted runner (push)
 
 on:
+  workflow_run:
+    workflows: ["Self-hosted runner (push-caller)"]
+    branches: ["main"]
+    types: [completed]
   push:
     branches:
-      - main
       - ci_*
       - ci-*
     paths:
@@ -20,37 +23,68 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   PYTEST_TIMEOUT: 60
+  TF_FORCE_GPU_ALLOW_GROWTH: true
+  RUN_PT_TF_CROSS_TESTS: 1
 
 jobs:
-  run_tests_torch_gpu:
-    runs-on: [self-hosted, docker-gpu, single-gpu]
-    container:
-      image: pytorch/pytorch:1.9.0-cuda11.1-cudnn8-runtime
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+  setup:
+    name: Setup
+    runs-on: ubuntu-latest
+    outputs:
+      matrix: ${{ steps.set-matrix.outputs.matrix }}
+      test_map: ${{ steps.set-matrix.outputs.test_map }}
     steps:
-      - name: Install dependencies
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # `CI_BRANCH_PUSH`: The branch name from the push event
+        # `CI_BRANCH_WORKFLOW_RUN`: The name of the branch on which this workflow is triggered by `workflow_run` event
+        # `CI_BRANCH`: The non-empty branch name from the above two (one and only one of them is empty)
+        # `CI_SHA_PUSH`: The commit SHA from the push event
+        # `CI_SHA_WORKFLOW_RUN`: The commit SHA that triggers this workflow by `workflow_run` event
+        # `CI_SHA`: The non-empty commit SHA from the above two (one and only one of them is empty)
         run: |
-          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git
-          apt install -y libsndfile1-dev espeak-ng
-          pip install --upgrade pip
-          pip install .[sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
 
-      - name: Launcher docker
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
+      - name: Checkout transformers
         uses: actions/checkout@v2
         with:
           fetch-depth: 2
 
-      - name: NVIDIA-SMI
+      - name: Update clone using environment variables
         run: |
-          nvidia-smi
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
 
-      - name: Are GPUs recognized by our DL frameworks
+      - name: Cleanup
         run: |
-          utils/print_env_pt.py
+          rm -rf tests/__pycache__
+          rm -rf tests/models/__pycache__
+          rm -rf reports
 
       - name: Fetch the tests to run
+        # TODO: add `git-python` in the docker images
         run: |
+          pip install --upgrade git-python
           python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
 
       - name: Report fetched tests
@@ -59,441 +93,445 @@ jobs:
           name: test_fetched
           path: test_preparation.txt
 
-      - name: Run all non-slow tests on GPU
+      - id: set-matrix
+        name: Organize tests into models
+        # The `keys` is used as GitHub actions matrix for jobs, i.e. `models/bert`, `tokenization`, `pipeline`, etc.
+        # The `test_map` is used to get the actual identified test files under each key.
+        # If no test to run (so no `test_map.json` file), create a dummy map (empty matrix will fail)
         run: |
-          if [ -f test_list.txt ]; then
-            python -m pytest -n 2 --dist=loadfile -v --make-reports=tests_torch_gpu $(cat test_list.txt)
+          if [ -f test_map.json ]; then
+              keys=$(python3 -c 'import json; fp = open("test_map.json"); test_map = json.load(fp); fp.close(); d = list(test_map.keys()); print(d)')
+              test_map=$(python3 -c 'import json; fp = open("test_map.json"); test_map = json.load(fp); fp.close(); print(test_map)')
+          else
+              keys=$(python3 -c 'keys = ["dummy"]; print(keys)')
+              test_map=$(python3 -c 'test_map = {"dummy": []}; print(test_map)')
           fi
+          echo $keys
+          echo $test_map
+          echo "::set-output name=matrix::$keys"
+          echo "::set-output name=test_map::$test_map"
+
+  run_check_runners:
+    name: Check Runners
+    needs: setup
+    strategy:
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
+    container:
+      image: huggingface/transformers-all-latest-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+  run_tests_single_gpu:
+    name: Model tests
+    needs: [setup, run_check_runners]
+    # `dummy` means there is no test to run
+    if: contains(fromJson(needs.setup.outputs.matrix), 'dummy') != true
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [single-gpu]
+    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
+    container:
+      image: huggingface/transformers-all-latest-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
+
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
+      - name: Update clone using environment variables
+        working-directory: /transformers
+        run: |
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
+
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          echo "${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
+
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
+      - name: Run all non-slow selected tests on GPU
+        working-directory: /transformers
+        run: |
+          python3 -m pytest -n 2 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} ${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}
 
       - name: Failure short reports
         if: ${{ failure() }}
-        run: cat reports/tests_torch_gpu/failures_short.txt
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_all_tests_torch_gpu_test_reports
-          path: reports
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
-#  run_tests_flax_gpu:
-#    runs-on: [self-hosted, docker-gpu-test, single-gpu]
-#    container:
-#      image: tensorflow/tensorflow:2.4.1-gpu
-#      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-#    steps:
-#      - name: Set up Python 3.7
-#        uses: actions/setup-python@v2
-#        with:
-#          python-version: 3.7
-#
-#      - name: Install dependencies
-#        run: |
-#          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
-#          pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html
-#          pip install --upgrade pip
-#          pip install .[sklearn,testing,sentencepiece,flax,flax-speech,vision]
-#
-#      - name: Launcher docker
-#        uses: actions/checkout@v2
-#        with:
-#          fetch-depth: 2
-#
-#      - name: NVIDIA-SMI
-#        continue-on-error: true
-#        run: |
-#          nvidia-smi
-#
-#      - name: Are GPUs recognized by our DL frameworks
-#        run: |
-#          python -c "from jax.lib import xla_bridge; print('GPU available:', xla_bridge.get_backend().platform)"
-#          python -c "import jax; print('Number of GPUs available:', len(jax.local_devices()))"
-#
-#      - name: Fetch the tests to run
-#        run: |
-#          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-#
-#      - name: Report fetched tests
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: test_fetched
-#          path: test_preparation.txt
-#
-#      - name: Run all non-slow tests on GPU
-#        run: |
-#          if [ -f test_list.txt ]; then
-#            python -m pytest -n 2 --dist=loadfile -v --make-reports=tests_flax_gpu $(cat test_list.txt)
-#          fi
-#
-#      - name: Failure short reports
-#        if: ${{ failure() }}
-#        run: cat reports/tests_flax_gpu/failures_short.txt
-#
-#      - name: Test suite reports artifacts
-#        if: ${{ always() }}
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: run_all_tests_flax_gpu_test_reports
-#          path: reports
-#
-#  run_tests_tf_gpu:
-#    runs-on: [self-hosted, docker-gpu, single-gpu]
-#    timeout-minutes: 120
-#    container:
-#      image: tensorflow/tensorflow:2.4.1-gpu
-#      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-#    steps:
-#      - name: Install dependencies
-#        run: |
-#          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
-#          pip install --upgrade pip
-#          pip install .[sklearn,testing,onnxruntime,sentencepiece,tf-speech]
-#          pip install https://github.com/kpu/kenlm/archive/master.zip
-#
-#      - name: Launcher docker
-#        uses: actions/checkout@v2
-#        with:
-#          fetch-depth: 2
-#
-#      - name: NVIDIA-SMI
-#        run: |
-#          nvidia-smi
-#
-#      - name: Are GPUs recognized by our DL frameworks
-#        run: |
-#          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-#          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
-#
-#      - name: Fetch the tests to run
-#        run: |
-#          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-#
-#      - name: Report fetched tests
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: test_fetched
-#          path: test_preparation.txt
-#
-#      - name: Run all non-slow tests on GPU
-#        env:
-#          TF_NUM_INTRAOP_THREADS: 8
-#          TF_NUM_INTEROP_THREADS: 1
-#        run: |
-#          if [ -f test_list.txt ]; then
-#            python -m pytest -n 1 --dist=loadfile --make-reports=tests_tf_gpu $(cat test_list.txt)
-#          fi
-#
-#      - name: Failure short reports
-#        if: ${{ failure() }}
-#        run: cat reports/tests_tf_gpu/failures_short.txt
-#
-#      - name: Test suite reports artifacts
-#        if: ${{ always() }}
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: run_all_tests_tf_gpu_test_reports
-#          path: reports
-
-
-  run_tests_torch_multi_gpu:
-    runs-on: [self-hosted, docker-gpu, multi-gpu]
+  run_tests_multi_gpu:
+    name: Model tests
+    needs: [setup, run_check_runners]
+    # `dummy` means there is no test to run
+    if: contains(fromJson(needs.setup.outputs.matrix), 'dummy') != true
+    strategy:
+      fail-fast: false
+      matrix:
+        folders: ${{ fromJson(needs.setup.outputs.matrix) }}
+        machine_type: [multi-gpu]
+    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
     container:
-      image: pytorch/pytorch:1.9.0-cuda11.1-cudnn8-runtime
+      image: huggingface/transformers-all-latest-gpu
       options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Install dependencies
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
         run: |
-          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
-          apt install -y libsndfile1-dev espeak-ng
-          pip install --upgrade pip
-          pip install .[sklearn,testing,onnxruntime,sentencepiece,torch-speech,vision,timm]
-          pip install https://github.com/kpu/kenlm/archive/master.zip
-      - name: Launcher docker
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 2
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
+
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
+      - name: Update clone using environment variables
+        working-directory: /transformers
+        run: |
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
+
+      - name: Echo folder ${{ matrix.folders }}
+        shell: bash
+        # For folders like `models/bert`, set an env. var. (`matrix_folders`) to `models_bert`, which will be used to
+        # set the artifact folder names (because the character `/` is not allowed).
+        run: |
+          echo "${{ matrix.folders }}"
+          echo "${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}"
+          matrix_folders=${{ matrix.folders }}
+          matrix_folders=${matrix_folders/'models/'/'models_'}
+          echo "$matrix_folders"
+          echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
 
       - name: NVIDIA-SMI
-        continue-on-error: true
         run: |
           nvidia-smi
 
-      - name: Are GPUs recognized by our DL frameworks
+      - name: Environment
+        working-directory: /transformers
         run: |
-          utils/print_env_pt.py
+          python3 utils/print_env.py
 
-      - name: Fetch the tests to run
-        run: |
-          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-
-      - name: Report fetched tests
-        uses: actions/upload-artifact@v2
-        with:
-          name: test_fetched
-          path: test_preparation.txt
-
-      - name: Run all non-slow tests on GPU
+      - name: Run all non-slow selected tests on GPU
         env:
           MKL_SERVICE_FORCE_INTEL: 1
+        working-directory: /transformers
         run: |
-          if [ -f test_list.txt ]; then
-            python -m pytest -n 2 --dist=loadfile -v --make-reports=tests_torch_multi_gpu $(cat test_list.txt)
-          fi
+          python3 -m pytest -n 2 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} ${{ fromJson(needs.setup.outputs.test_map)[matrix.folders] }}
 
       - name: Failure short reports
         if: ${{ failure() }}
-        run: cat reports/tests_torch_multi_gpu/failures_short.txt
+        continue-on-error: true
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_all_tests_torch_multi_gpu_test_reports
-          path: reports
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
-#  run_tests_flax_multi_gpu:
-#    runs-on: [self-hosted, docker-gpu, multi-gpu]
-#    container:
-#      image: tensorflow/tensorflow:2.4.1-gpu
-#      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-#    steps:
-#      - name: Install dependencies
-#        run: |
-#          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
-#          pip install --upgrade "jax[cuda111]" -f https://storage.googleapis.com/jax-releases/jax_releases.html
-#          pip install --upgrade pip
-#          pip install .[sklearn,testing,sentencepiece,flax,flax-speech,vision]
-#          pip install https://github.com/kpu/kenlm/archive/master.zip
-#
-#      - name: Launcher docker
-#        uses: actions/checkout@v2
-#        with:
-#          fetch-depth: 2
-#
-#      - name: NVIDIA-SMI
-#        continue-on-error: true
-#        run: |
-#          nvidia-smi
-#
-#      - name: Are GPUs recognized by our DL frameworks
-#        run: |
-#          python -c "from jax.lib import xla_bridge; print('GPU available:', xla_bridge.get_backend().platform)"
-#          python -c "import jax; print('Number of GPUs available:', len(jax.local_devices()))"
-#
-#      - name: Fetch the tests to run
-#        run: |
-#          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-#
-#      - name: Report fetched tests
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: test_fetched
-#          path: test_preparation.txt
-#
-#      - name: Run all non-slow tests on GPU
-#        run: |
-#          if [ -f test_list.txt ]; then
-#            python -m pytest -n 2 --dist=loadfile -v --make-reports=tests_flax_multi_gpu $(cat test_list.txt)
-#          fi
-#
-#      - name: Failure short reports
-#        if: ${{ failure() }}
-#        run: cat reports/tests_flax_multi_gpu/failures_short.txt
-#
-#      - name: Test suite reports artifacts
-#        if: ${{ always() }}
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: run_all_tests_flax_multi_gpu_test_reports
-#          path: reports
-
-#  run_tests_tf_multi_gpu:
-#    runs-on: [self-hosted, docker-gpu, multi-gpu]
-#    timeout-minutes: 120
-#    container:
-#      image: tensorflow/tensorflow:2.4.1-gpu
-#      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-#    steps:
-#      - name: Install dependencies
-#        run: |
-#          apt -y update && apt install -y software-properties-common && apt -y update && add-apt-repository -y ppa:git-core/ppa && apt -y update && apt install -y git espeak-ng
-#          pip install --upgrade pip
-#          pip install .[sklearn,testing,onnxruntime,sentencepiece,tf-speech]
-#          pip install https://github.com/kpu/kenlm/archive/master.zip
-#
-#      - name: Launcher docker
-#        uses: actions/checkout@v2
-#        with:
-#          fetch-depth: 2
-#
-#      - name: NVIDIA-SMI
-#        run: |
-#          nvidia-smi
-#
-#      - name: Are GPUs recognized by our DL frameworks
-#        run: |
-#          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-#          TF_CPP_MIN_LOG_LEVEL=3 python -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
-#
-#      - name: Fetch the tests to run
-#        run: |
-#          python utils/tests_fetcher.py --diff_with_last_commit | tee test_preparation.txt
-#
-#      - name: Report fetched tests
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: test_fetched
-#          path: test_preparation.txt
-#
-#      - name: Run all non-slow tests on GPU
-#        env:
-#          TF_NUM_INTRAOP_THREADS: 8
-#          TF_NUM_INTEROP_THREADS: 1
-#        run: |
-#          if [ -f test_list.txt ]; then
-#            python -m pytest -n 1 --dist=loadfile --make-reports=tests_tf_multi_gpu $(cat test_list.txt)
-#          fi
-#
-#      - name: Failure short reports
-#        if: ${{ failure() }}
-#        run: cat reports/tests_tf_multi_gpu/failures_short.txt
-#
-#      - name: Test suite reports artifacts
-#        if: ${{ always() }}
-#        uses: actions/upload-artifact@v2
-#        with:
-#          name: run_all_tests_tf_multi_gpu_test_reports
-#          path: reports
-
-  run_tests_torch_cuda_extensions_gpu:
-    runs-on: [self-hosted, docker-gpu, single-gpu]
+  run_tests_torch_cuda_extensions_single_gpu:
+    name: Torch CUDA extension tests
+    needs: [setup, run_check_runners]
+    if: contains(fromJson(needs.setup.outputs.matrix), 'deepspeed') || contains(fromJson(needs.setup.outputs.matrix), 'extended')
+    strategy:
+      fail-fast: false
+      matrix:
+        machine_type: [single-gpu]
+    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
     container:
-      image: nvcr.io/nvidia/pytorch:21.03-py3
+      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 2
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
+
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
+      - name: Update clone using environment variables
+        working-directory: /workspace/transformers
+        run: |
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
+
+      - name: Remove cached torch extensions
+        run: rm -rf /github/home/.cache/torch_extensions/
+
+      # To avoid unknown test failures
+      - name: Pre build DeepSpeed *again*
+        working-directory: /workspace
+        run: |
+          python3 -m pip uninstall -y deepspeed
+          DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
 
       - name: NVIDIA-SMI
         run: |
           nvidia-smi
 
-      - name: Install dependencies
+      - name: Environment
+        working-directory: /workspace/transformers
         run: |
-          apt -y update && apt install -y libaio-dev
-          pip install --upgrade pip
-          pip install .[deepspeed-testing]
+          python utils/print_env.py
 
-      - name: Are GPUs recognized by our DL frameworks
+      - name: Run all non-slow selected tests on GPU
+        working-directory: /workspace/transformers
+        # TODO: Here we pass all tests in the 2 folders for simplicity. It's better to pass only the identified tests.
         run: |
-          utils/print_env_pt.py
-
-      - name: Fetch the tests to run
-        run: |
-          python utils/tests_fetcher.py --diff_with_last_commit --filters tests/deepspeed tests/extended | tee test_preparation.txt
-
-      - name: Report fetched tests
-        uses: actions/upload-artifact@v2
-        with:
-          name: test_fetched
-          path: test_preparation.txt
-
-      - name: Run all tests on GPU
-        run: |
-          if [ -f test_list.txt ]; then
-            python -m pytest -n 1 --dist=loadfile -v --make-reports=tests_torch_cuda_extensions_gpu $(cat test_list.txt)
-          fi
+          python -m pytest -n 1 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
 
       - name: Failure short reports
         if: ${{ failure() }}
-        run: cat reports/tests_torch_cuda_extensions_gpu/failures_short.txt
+        continue-on-error: true
+        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_tests_torch_cuda_extensions_gpu_test_reports
-          path: reports
+          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
+          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
 
   run_tests_torch_cuda_extensions_multi_gpu:
-    runs-on: [self-hosted, docker-gpu, multi-gpu]
+    name: Torch CUDA extension tests
+    needs: [setup, run_check_runners]
+    if: contains(fromJson(needs.setup.outputs.matrix), 'deepspeed') || contains(fromJson(needs.setup.outputs.matrix), 'extended')
+    strategy:
+      fail-fast: false
+      matrix:
+        machine_type: [multi-gpu]
+    runs-on: [self-hosted, docker-gpu, '${{ matrix.machine_type }}']
     container:
-      image: nvcr.io/nvidia/pytorch:21.03-py3
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+      image: huggingface/transformers-pytorch-deepspeed-latest-gpu
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
     steps:
-      - name: Launcher docker
-        uses: actions/checkout@v2
-        with:
-          fetch-depth: 2
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
+
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
+      - name: Update clone using environment variables
+        working-directory: /workspace/transformers
+        run: |
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
+
+      - name: Remove cached torch extensions
+        run: rm -rf /github/home/.cache/torch_extensions/
+
+      # To avoid unknown test failures
+      - name: Pre build DeepSpeed *again*
+        working-directory: /workspace
+        run: |
+          python3 -m pip uninstall -y deepspeed
+          DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
 
       - name: NVIDIA-SMI
-        continue-on-error: true
         run: |
           nvidia-smi
 
-      - name: Install dependencies
+      - name: Environment
+        working-directory: /workspace/transformers
         run: |
-          apt -y update && apt install -y libaio-dev
-          pip install --upgrade pip
-          rm -rf ~/.cache/torch_extensions/ # shared between conflicting builds
-          pip install .[testing,deepspeed,fairscale]
+          python utils/print_env.py
 
-      - name: Are GPUs recognized by our DL frameworks
+      - name: Run all non-slow selected tests on GPU
+        working-directory: /workspace/transformers
+        # TODO: Here we pass all tests in the 2 folders for simplicity. It's better to pass only the identified tests.
         run: |
-          utils/print_env_pt.py
-
-      - name: Fetch the tests to run
-        run: |
-          python utils/tests_fetcher.py --diff_with_last_commit --filters tests/deepspeed tests/extended | tee test_preparation.txt
-
-      - name: Report fetched tests
-        uses: actions/upload-artifact@v2
-        with:
-          name: test_fetched
-          path: test_preparation.txt
-
-      - name: Run all tests on GPU
-        run: |
-          if [ -f test_list.txt ]; then
-            python -m pytest -n 1 --dist=loadfile -v --make-reports=tests_torch_cuda_extensions_multi_gpu $(cat test_list.txt)
-          fi
+          python -m pytest -n 1 --dist=loadfile -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
 
       - name: Failure short reports
         if: ${{ failure() }}
-        run: cat reports/tests_torch_cuda_extensions_multi_gpu/failures_short.txt
+        continue-on-error: true
+        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_tests_torch_cuda_extensions_multi_gpu_test_reports
-          path: reports
-
+          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
+          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
 
   send_results:
     name: Send results to webhook
     runs-on: ubuntu-latest
     if: always()
     needs: [
-        run_tests_torch_gpu,
-#        run_tests_tf_gpu,
-        run_tests_torch_multi_gpu,
-#        run_tests_tf_multi_gpu,
-        run_tests_torch_cuda_extensions_gpu,
+        setup,
+        run_check_runners,
+        run_tests_single_gpu,
+        run_tests_multi_gpu,
+        run_tests_torch_cuda_extensions_single_gpu,
         run_tests_torch_cuda_extensions_multi_gpu
     ]
     steps:
+      - name: Preliminary job status
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          echo "Setup status: ${{ needs.setup.result }}"
+          echo "Runner status: ${{ needs.run_check_runners.result }}"
+
+      # Necessary to get the correct branch name and commit SHA for `workflow_run` event
+      # We also take into account the `push` event (we might want to test some changes in a branch)
+      - name: Prepare custom environment variables
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          CI_BRANCH_PUSH=${{ github.event.ref }}
+          CI_BRANCH_PUSH=${CI_BRANCH_PUSH/'refs/heads/'/''}
+          CI_BRANCH_WORKFLOW_RUN=${{ github.event.workflow_run.head_branch }}
+          CI_SHA_PUSH=${{ github.event.head_commit.id }}
+          CI_SHA_WORKFLOW_RUN=${{ github.event.workflow_run.head_sha }}
+          echo $CI_BRANCH_PUSH
+          echo $CI_BRANCH_WORKFLOW_RUN
+          echo $CI_SHA_PUSH
+          echo $CI_SHA_WORKFLOW_RUN
+          [[ ! -z "$CI_BRANCH_PUSH" ]] && echo "CI_BRANCH=$CI_BRANCH_PUSH" >> $GITHUB_ENV || echo "CI_BRANCH=$CI_BRANCH_WORKFLOW_RUN" >> $GITHUB_ENV
+          [[ ! -z "$CI_SHA_PUSH" ]] && echo "CI_SHA=$CI_SHA_PUSH" >> $GITHUB_ENV || echo "CI_SHA=$CI_SHA_WORKFLOW_RUN" >> $GITHUB_ENV
+
+      - name: print environment variables
+        run: |
+          echo "env.CI_BRANCH = ${{ env.CI_BRANCH }}"
+          echo "env.CI_SHA = ${{ env.CI_SHA }}"
+
       - uses: actions/checkout@v2
 
-      - uses: actions/download-artifact@v2
+      - name: Update clone using environment variables
+        run: |
+          echo "original branch = $(git branch --show-current)"
+          git fetch && git checkout ${{ env.CI_BRANCH }}
+          echo "updated branch = $(git branch --show-current)"
+          git checkout ${{ env.CI_SHA }}
+          echo "log = $(git log -n 1)"
 
+      - uses: actions/download-artifact@v2
       - name: Send message to Slack
         env:
           CI_SLACK_BOT_TOKEN: ${{ secrets.CI_SLACK_BOT_TOKEN }}
           CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
+          CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
+          CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
+          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
+          CI_EVENT: push
+          CI_TITLE_PUSH: ${{ github.event.head_commit.message }}
+          CI_TITLE_WORKFLOW_RUN: ${{ github.event.workflow_run.head_commit.message }}
+          CI_SHA: ${{ env.CI_SHA }}
+          SETUP_STATUS: ${{ needs.setup.result }}
+          RUNNER_STATUS: ${{ needs.run_check_runners.result }}
 
+        # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
+        # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
         run: |
           pip install slack_sdk
-          python utils/notification_service_deprecated.py push
+          python utils/notification_service.py "${{ needs.setup.outputs.matrix }}"

.github/workflows/self-scheduled.yml

@@ -22,12 +22,27 @@ env:
   RUN_PT_TF_CROSS_TESTS: 1
 
 jobs:
-  setup:
-    name: Setup
+  run_check_runners:
+    name: Check Runners
     strategy:
       matrix:
-        machines: [multi-gpu-docker, single-gpu-docker]
-    runs-on: ${{ matrix.machines }}
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    container:
+      image: huggingface/transformers-all-latest-gpu
+      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    steps:
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+  setup:
+    name: Setup
+    needs: run_check_runners
+    strategy:
+      matrix:
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
     container:
       image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
@@ -56,25 +71,18 @@ jobs:
         run: |
           nvidia-smi
 
-      - name: GPU visibility
-        working-directory: /transformers
-        run: |
-          utils/print_env_pt.py
-          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('TF GPUs available:', bool(tf.config.list_physical_devices('GPU')))"
-          TF_CPP_MIN_LOG_LEVEL=3 python3 -c "import tensorflow as tf; print('Number of TF GPUs available:', len(tf.config.list_physical_devices('GPU')))"
-
   run_tests_single_gpu:
     name: Model tests
     strategy:
       fail-fast: false
       matrix:
         folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machines: [single-gpu-docker]
-    runs-on: ${{ matrix.machines }}
+        machine_type: [single-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
     container:
       image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
+    needs: [run_check_runners, setup]
     steps:
       - name: Echo folder ${{ matrix.folders }}
         shell: bash
@@ -83,7 +91,6 @@ jobs:
         run: |
           echo "${{ matrix.folders }}"
           matrix_folders=${{ matrix.folders }}
-          echo "$matrix_folders"
           matrix_folders=${matrix_folders/'models/'/'models_'}
           echo "$matrix_folders"
           echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
@@ -92,21 +99,30 @@ jobs:
         working-directory: /transformers
         run: git fetch && git checkout ${{ github.sha }}
 
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
       - name: Run all tests on GPU
         working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machines }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: ${{ matrix.machines }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
   run_tests_multi_gpu:
     name: Model tests
@@ -114,12 +130,12 @@ jobs:
       fail-fast: false
       matrix:
         folders: ${{ fromJson(needs.setup.outputs.matrix) }}
-        machines: [multi-gpu-docker]
-    runs-on: ${{ matrix.machines }}
+        machine_type: [multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
     container:
       image: huggingface/transformers-all-latest-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [run_check_runners, setup]
     steps:
       - name: Echo folder ${{ matrix.folders }}
         shell: bash
@@ -128,7 +144,6 @@ jobs:
         run: |
           echo "${{ matrix.folders }}"
           matrix_folders=${{ matrix.folders }}
-          echo "$matrix_folders"
           matrix_folders=${matrix_folders/'models/'/'models_'}
           echo "$matrix_folders"
           echo "matrix_folders=$matrix_folders" >> $GITHUB_ENV
@@ -137,21 +152,30 @@ jobs:
         working-directory: /transformers
         run: git fetch && git checkout ${{ github.sha }}
 
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
       - name: Run all tests on GPU
         working-directory: /transformers
-        run: python3 -m pytest -v --make-reports=${{ matrix.machines }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
+        run: python3 -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }} tests/${{ matrix.folders }}
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: ${{ matrix.machines }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
-          path: /transformers/reports/${{ matrix.machines }}_tests_gpu_${{ matrix.folders }}
+          name: ${{ matrix.machine_type }}_run_all_tests_gpu_${{ env.matrix_folders }}_test_reports
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_gpu_${{ matrix.folders }}
 
   run_examples_gpu:
     name: Examples directory
@@ -159,109 +183,136 @@ jobs:
     container:
       image: huggingface/transformers-all-latest-gpu
       options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
+    needs: [run_check_runners, setup]
     steps:
       - name: Update clone
         working-directory: /transformers
         run: git fetch && git checkout ${{ github.sha }}
 
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
       - name: Run examples tests on GPU
         working-directory: /transformers
         run: |
           pip install -r examples/pytorch/_tests_requirements.txt
-          python3 -m pytest -v --make-reports=examples_gpu examples/pytorch
+          python3 -m pytest -v --make-reports=single-gpu_examples_gpu examples/pytorch
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/examples_gpu/failures_short.txt
+        run: cat /transformers/reports/single-gpu_examples_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: run_examples_gpu
-          path: /transformers/reports/examples_gpu
+          name: single-gpu_run_examples_gpu
+          path: /transformers/reports/single-gpu_examples_gpu
 
   run_pipelines_torch_gpu:
     name: PyTorch pipelines
     strategy:
       fail-fast: false
       matrix:
-        machines: [multi-gpu-docker, single-gpu-docker]
-    runs-on: ${{ matrix.machines }}
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
     container:
       image: huggingface/transformers-pytorch-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [run_check_runners, setup]
     steps:
       - name: Update clone
         working-directory: /transformers
         run: git fetch && git checkout ${{ github.sha }}
 
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
       - name: Run all pipeline tests on GPU
         working-directory: /transformers
         env:
           RUN_PIPELINE_TESTS: yes
         run: |
-          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machines }}_tests_torch_pipeline_gpu tests
+          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machine_type }}_tests_torch_pipeline_gpu tests
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /transformers/reports/${{ matrix.machines }}_tests_torch_pipeline_gpu/failures_short.txt
+        run: cat /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: ${{ matrix.machines }}_run_tests_torch_pipeline_gpu
-          path: /transformers/reports/${{ matrix.machines }}_tests_torch_pipeline_gpu
+          name: ${{ matrix.machine_type }}_run_tests_torch_pipeline_gpu
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_torch_pipeline_gpu
 
   run_pipelines_tf_gpu:
     name: TensorFlow pipelines
     strategy:
       fail-fast: false
       matrix:
-        machines: [multi-gpu-docker, single-gpu-docker]
-    runs-on: ${{ matrix.machines }}
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
     container:
       image: huggingface/transformers-tensorflow-gpu
-      options: --gpus 0 --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
-    needs: setup
+      options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
+    needs: [run_check_runners, setup]
     steps:
       - name: Update clone
         working-directory: /transformers
         run: |
           git fetch && git checkout ${{ github.sha }}
 
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /transformers
+        run: |
+          python3 utils/print_env.py
+
       - name: Run all pipeline tests on GPU
         working-directory: /transformers
         env:
           RUN_PIPELINE_TESTS: yes
         run: |
-          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machines }}_tests_tf_pipeline_gpu tests
+          python3 -m pytest -n 1 -v --dist=loadfile -m is_pipeline_test --make-reports=${{ matrix.machine_type }}_tests_tf_pipeline_gpu tests
 
       - name: Failure short reports
         if: ${{ always() }}
         run: |
-          cat /transformers/reports/${{ matrix.machines }}_tests_tf_pipeline_gpu/failures_short.txt
+          cat /transformers/reports/${{ matrix.machine_type }}_tests_tf_pipeline_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: ${{ matrix.machines }}_run_tests_tf_pipeline_gpu
-          path: /transformers/reports/${{ matrix.machines }}_tests_tf_pipeline_gpu
+          name: ${{ matrix.machine_type }}_run_tests_tf_pipeline_gpu
+          path: /transformers/reports/${{ matrix.machine_type }}_tests_tf_pipeline_gpu
 
   run_all_tests_torch_cuda_extensions_gpu:
     name: Torch CUDA extension tests
     strategy:
       fail-fast: false
       matrix:
-        machines: [multi-gpu-docker, single-gpu-docker]
-    runs-on: ${{ matrix.machines }}
-    needs: setup
+        machine_type: [single-gpu, multi-gpu]
+    runs-on: ${{ format('{0}-{1}', matrix.machine_type, 'docker') }}
+    needs: [run_check_runners, setup]
     container:
       image: huggingface/transformers-pytorch-deepspeed-latest-gpu
       options: --gpus all --shm-size "16gb" --ipc host -v /mnt/cache/.cache/huggingface:/mnt/cache/
@@ -270,38 +321,64 @@ jobs:
         working-directory: /workspace/transformers
         run: git fetch && git checkout ${{ github.sha }}
 
-      - name: Re-compile DeepSpeed
+      - name: Remove cached torch extensions
+        run: rm -rf /github/home/.cache/torch_extensions/
+
+      # To avoid unknown test failures
+      - name: Pre build DeepSpeed *again*
         working-directory: /workspace
         run: |
-          pip install deepspeed # installs the deps correctly
-          rm -rf DeepSpeed
-          git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build
-          DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install -e . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
+          python3 -m pip uninstall -y deepspeed
+          DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check
+
+      - name: NVIDIA-SMI
+        run: |
+          nvidia-smi
+
+      - name: Environment
+        working-directory: /workspace/transformers
+        run: |
+          python utils/print_env.py
 
       - name: Run all tests on GPU
         working-directory: /workspace/transformers
         run: |
-          python -m pytest -v --make-reports=${{ matrix.machines }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
+          python -m pytest -v --make-reports=${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu tests/deepspeed tests/extended
 
       - name: Failure short reports
         if: ${{ failure() }}
         continue-on-error: true
-        run: cat /workspace/transformers/reports/${{ matrix.machines }}_tests_torch_cuda_extensions_gpu/failures_short.txt
+        run: cat /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu/failures_short.txt
 
       - name: Test suite reports artifacts
         if: ${{ always() }}
         uses: actions/upload-artifact@v2
         with:
-          name: ${{ matrix.machines }}_run_tests_torch_cuda_extensions_gpu_test_reports
-          path: /workspace/transformers/reports/${{ matrix.machines }}_tests_torch_cuda_extensions_gpu
-
+          name: ${{ matrix.machine_type }}_run_tests_torch_cuda_extensions_gpu_test_reports
+          path: /workspace/transformers/reports/${{ matrix.machine_type }}_tests_torch_cuda_extensions_gpu
 
   send_results:
     name: Send results to webhook
     runs-on: ubuntu-latest
     if: always()
-    needs: [setup, run_tests_single_gpu, run_tests_multi_gpu, run_examples_gpu, run_pipelines_tf_gpu, run_pipelines_torch_gpu, run_all_tests_torch_cuda_extensions_gpu]
+    needs: [
+      run_check_runners,
+      setup,
+      run_tests_single_gpu,
+      run_tests_multi_gpu,
+      run_examples_gpu,
+      run_pipelines_tf_gpu,
+      run_pipelines_torch_gpu,
+      run_all_tests_torch_cuda_extensions_gpu
+    ]
     steps:
+      - name: Preliminary job status
+        shell: bash
+        # For the meaning of these environment variables, see the job `Setup`
+        run: |
+          echo "Runner status: ${{ needs.run_check_runners.result }}"
+          echo "Setup status: ${{ needs.setup.result }}"
+
       - uses: actions/checkout@v2
       - uses: actions/download-artifact@v2
       - name: Send message to Slack
@@ -310,6 +387,10 @@ jobs:
           CI_SLACK_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID }}
           CI_SLACK_CHANNEL_ID_DAILY: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
           CI_SLACK_CHANNEL_DUMMY_TESTS: ${{ secrets.CI_SLACK_CHANNEL_DUMMY_TESTS }}
+          CI_SLACK_REPORT_CHANNEL_ID: ${{ secrets.CI_SLACK_CHANNEL_ID_DAILY }}
+          CI_EVENT: scheduled
+          SETUP_STATUS: ${{ needs.setup.result }}
+          RUNNER_STATUS: ${{ needs.run_check_runners.result }}
         # We pass `needs.setup.outputs.matrix` as the argument. A processing in `notification_service.py` to change
         # `models/bert` to `models_bert` is required, as the artifact names use `_` instead of `/`.
         run: |

.github/workflows/update_metdata.yml

@@ -21,7 +21,7 @@ jobs:
         id: cache
         with:
           path: ~/venv/
-          key: v2-metadata-${{ hashFiles('setup.py') }}
+          key: v3-metadata-${{ hashFiles('setup.py') }}
 
       - name: Create virtual environment on cache miss
         if: steps.cache.outputs.cache-hit != 'true'

CONTRIBUTING.md

@@ -128,7 +128,7 @@ You will need basic `git` proficiency to be able to contribute to
 manual. Type `git --help` in a shell and enjoy. If you prefer books, [Pro
 Git](https://git-scm.com/book/en/v2) is a very good reference.
 
-Follow these steps to start contributing:
+Follow these steps to start contributing ([supported Python versions](https://github.com/huggingface/transformers/blob/main/setup.py#L426)):
 
 1. Fork the [repository](https://github.com/huggingface/transformers) by
    clicking on the 'Fork' button on the repository's page. This creates a copy of the code
@@ -171,6 +171,14 @@ Follow these steps to start contributing:
 
    If you have already cloned that repo, you might need to `git pull` to get the most recent changes in the `datasets`
    library.
+   
+   Depending on your OS, you might need to install some external libraries, as well, if the `pip` installation fails.
+   
+   For macOS, you will likely need [MeCab](https://taku910.github.io/mecab/), which can be installed from Homebrew:
+   
+   ```bash
+   brew install mecab
+   ```
 
 5. Develop the features on your branch.
 

Makefile

@@ -9,7 +9,7 @@ modified_only_fixup:
 	$(eval modified_py_files := $(shell python utils/get_modified_files.py $(check_dirs)))
 	@if test -n "$(modified_py_files)"; then \
 		echo "Checking/fixing $(modified_py_files)"; \
-		black $(modified_py_files); \
+		black --preview $(modified_py_files); \
 		isort $(modified_py_files); \
 		flake8 $(modified_py_files); \
 	else \
@@ -45,22 +45,26 @@ repo-consistency:
 # this target runs checks on all files
 
 quality:
-	black --check $(check_dirs)
+	black --check --preview $(check_dirs)
 	isort --check-only $(check_dirs)
 	python utils/custom_init_isort.py --check_only
+	python utils/sort_auto_mappings.py --check_only
 	flake8 $(check_dirs)
 	doc-builder style src/transformers docs/source --max_len 119 --check_only --path_to_docs docs/source
+	python utils/check_doc_toc.py
 
 # Format source code automatically and check is there are any problems left that need manual fixing
 
 extra_style_checks:
 	python utils/custom_init_isort.py
+	python utils/sort_auto_mappings.py
 	doc-builder style src/transformers docs/source --max_len 119 --path_to_docs docs/source
+	python utils/check_doc_toc.py --fix_and_overwrite
 
 # this target runs checks on all files and potentially modifies some of them
 
 style:
-	black $(check_dirs)
+	black --preview $(check_dirs)
 	isort $(check_dirs)
 	${MAKE} autogenerate_code
 	${MAKE} extra_style_checks

README.md

@@ -87,12 +87,16 @@ Here are a few examples:
 In Computer Vision:
 - [Image classification with ViT](https://huggingface.co/google/vit-base-patch16-224)
 - [Object Detection with DETR](https://huggingface.co/facebook/detr-resnet-50)
-- [Image Segmentation with DETR](https://huggingface.co/facebook/detr-resnet-50-panoptic)
+- [Semantic Segmentation with SegFormer](https://huggingface.co/nvidia/segformer-b0-finetuned-ade-512-512)
+- [Panoptic Segmentation with DETR](https://huggingface.co/facebook/detr-resnet-50-panoptic)
 
 In Audio:
 - [Automatic Speech Recognition with Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base-960h)
 - [Keyword Spotting with Wav2Vec2](https://huggingface.co/superb/wav2vec2-base-superb-ks)
 
+In Multimodal tasks:
+- [Visual Question Answering with ViLT](https://huggingface.co/dandelin/vilt-b32-finetuned-vqa)
+
 **[Write With Transformer](https://transformer.huggingface.co)**, built by the Hugging Face team, is the official demo of this repo’s text generation capabilities.
 
 ## If you are looking for custom support from the Hugging Face team
@@ -116,24 +120,48 @@ To immediately use a model on a given input (text, image, audio, ...), we provid
 
 The second line of code downloads and caches the pretrained model used by the pipeline, while the third evaluates it on the given text. Here the answer is "positive" with a confidence of 99.97%.
 
-Many NLP tasks have a pre-trained `pipeline` ready to go. For example, we can easily extract question answers given context:
+Many tasks have a pre-trained `pipeline` ready to go, in NLP but also in computer vision and speech. For example, we can easily extract detected objects in an image:
 
 ``` python
+>>> import requests
+>>> from PIL import Image
 >>> from transformers import pipeline
 
-# Allocate a pipeline for question-answering
->>> question_answerer = pipeline('question-answering')
->>> question_answerer({
-...     'question': 'What is the name of the repository ?',
-...     'context': 'Pipeline has been included in the huggingface/transformers repository'
-... })
-{'score': 0.30970096588134766, 'start': 34, 'end': 58, 'answer': 'huggingface/transformers'}
+# Download an image with cute cats
+>>> url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample.png"
+>>> image_data = requests.get(url, stream=True).raw
+>>> image = Image.open(image_data)
 
+# Allocate a pipeline for object detection
+>>> object_detector = pipeline('object_detection')
+>>> object_detector(image)
+[{'score': 0.9982201457023621,
+  'label': 'remote',
+  'box': {'xmin': 40, 'ymin': 70, 'xmax': 175, 'ymax': 117}},
+ {'score': 0.9960021376609802,
+  'label': 'remote',
+  'box': {'xmin': 333, 'ymin': 72, 'xmax': 368, 'ymax': 187}},
+ {'score': 0.9954745173454285,
+  'label': 'couch',
+  'box': {'xmin': 0, 'ymin': 1, 'xmax': 639, 'ymax': 473}},
+ {'score': 0.9988006353378296,
+  'label': 'cat',
+  'box': {'xmin': 13, 'ymin': 52, 'xmax': 314, 'ymax': 470}},
+ {'score': 0.9986783862113953,
+  'label': 'cat',
+  'box': {'xmin': 345, 'ymin': 23, 'xmax': 640, 'ymax': 368}}]
 ```
 
-In addition to the answer, the pretrained model used here returned its confidence score, along with the start position and end position of the answer in the tokenized sentence. You can learn more about the tasks supported by the `pipeline` API in [this tutorial](https://huggingface.co/docs/transformers/task_summary).
+Here we get a list of objects detected in the image, with a box surrounding the object and a confidence score. Here is the original image on the right, with the predictions displayed on the left:
 
-To download and use any of the pretrained models on your given task, all it takes is three lines of code. Here is the PyTorch version:
+<h3 align="center">
+    <a><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample.png" width="400"></a>
+    <a><img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/coco_sample_post_processed.png" width="400"></a>
+</h3>
+
+You can learn more about the tasks supported by the `pipeline` API in [this tutorial](https://huggingface.co/docs/transformers/task_summary).
+
+In addition to `pipeline`, to download and use any of the pretrained models on your given task, all it takes is three lines of code. Here is the PyTorch version:
 ```python
 >>> from transformers import AutoTokenizer, AutoModel
 
@@ -143,6 +171,7 @@ To download and use any of the pretrained models on your given task, all it take
 >>> inputs = tokenizer("Hello world!", return_tensors="pt")
 >>> outputs = model(**inputs)
 ```
+
 And here is the equivalent code for TensorFlow:
 ```python
 >>> from transformers import AutoTokenizer, TFAutoModel
@@ -156,7 +185,7 @@ And here is the equivalent code for TensorFlow:
 
 The tokenizer is responsible for all the preprocessing the pretrained model expects, and can be called directly on a single string (as in the above examples) or a list. It will output a dictionary that you can use in downstream code or simply directly pass to your model using the ** argument unpacking operator.
 
-The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) or a [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) (depending on your backend) which you can use normally. [This tutorial](https://huggingface.co/docs/transformers/training) explains how to integrate such a model into a classic PyTorch or TensorFlow training loop, or how to use our `Trainer` API to quickly fine-tune on a new dataset.
+The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) or a [TensorFlow `tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) (depending on your backend) which you can use as usual. [This tutorial](https://huggingface.co/docs/transformers/training) explains how to integrate such a model into a classic PyTorch or TensorFlow training loop, or how to use our `Trainer` API to quickly fine-tune on a new dataset.
 
 ## Why should I use transformers?
 
@@ -169,7 +198,7 @@ The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/sta
 1. Lower compute costs, smaller carbon footprint:
     - Researchers can share trained models instead of always retraining.
     - Practitioners can reduce compute time and production costs.
-    - Dozens of architectures with over 20,000 pretrained models, some in more than 100 languages.
+    - Dozens of architectures with over 60,000 pretrained models across all modalities.
 
 1. Choose the right framework for every part of a model's lifetime:
     - Train state-of-the-art models in 3 lines of code.
@@ -184,7 +213,7 @@ The model itself is a regular [Pytorch `nn.Module`](https://pytorch.org/docs/sta
 ## Why shouldn't I use transformers?
 
 - This library is not a modular toolbox of building blocks for neural nets. The code in the model files is not refactored with additional abstractions on purpose, so that researchers can quickly iterate on each of the models without diving into additional abstractions/files.
-- The training API is not intended to work on any model but is optimized to work with the models provided by the library. For generic machine learning loops, you should use another library.
+- The training API is not intended to work on any model but is optimized to work with the models provided by the library. For generic machine learning loops, you should use another library (possibly, [Accelerate](https://huggingface.co/docs/accelerate)).
 - While we strive to present as many use cases as possible, the scripts in our [examples folder](https://github.com/huggingface/transformers/tree/main/examples) are just that: examples. It is expected that they won't work out-of-the box on your specific problem and that you will be required to change a few lines of code to adapt them to your needs.
 
 ## Installation
@@ -220,6 +249,8 @@ conda install -c huggingface transformers
 
 Follow the installation pages of Flax, PyTorch or TensorFlow to see how to install them with conda.
 
+> **_NOTE:_**  On Windows, you may be prompted to activate Developer Mode in order to benefit from caching. If this is not an option for you, please let us know in [this issue](https://github.com/huggingface/huggingface_hub/issues/1062).
+
 ## Model architectures
 
 **[All the model checkpoints](https://huggingface.co/models)** provided by 🤗 Transformers are seamlessly integrated from the huggingface.co [model hub](https://huggingface.co) where they are uploaded directly by [users](https://huggingface.co/users) and [organizations](https://huggingface.co/organizations).
@@ -234,78 +265,95 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[BARTpho](https://huggingface.co/docs/transformers/model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
 1. **[BEiT](https://huggingface.co/docs/transformers/model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
 1. **[BERT](https://huggingface.co/docs/transformers/model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
 1. **[BERT For Sequence Generation](https://huggingface.co/docs/transformers/model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[BERTweet](https://huggingface.co/docs/transformers/model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
 1. **[BigBird-Pegasus](https://huggingface.co/docs/transformers/model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
 1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
 1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
 1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
 1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
+1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
+1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DeiT](https://huggingface.co/docs/transformers/model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](https://huggingface.co/docs/transformers/model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
 1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
-1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval
-for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon
-Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
+1. **[Donut](https://huggingface.co/docs/transformers/model_doc/donut)** (from NAVER), released together with the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
+1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
+1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[ERNIE](https://huggingface.co/docs/transformers/model_doc/ernie)** (from Baidu) released with the paper [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/main/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
+1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
+1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
+1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
+1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
 1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
 1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
 1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
+1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
+1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
+1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
+1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
+1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[PEGASUS-X](https://huggingface.co/docs/transformers/model_doc/pegasus_x)** (from Google) released with the paper [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) by Jason Phang, Yao Zhao, and Peter J. Liu.
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
 1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
 1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
+1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
+1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
-1. **[RegNet](https://huggingface.co/docs/transformers/main/model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
 1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
 1. **[RoFormer](https://huggingface.co/docs/transformers/model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
@@ -315,34 +363,39 @@ Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/main/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
 1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
+1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
-1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER
-AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
+1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
 1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[VideoMAE](https://huggingface.co/docs/transformers/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
 1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
+1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
 1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[X-CLIP](https://huggingface.co/docs/transformers/model_doc/xclip)** (from Microsoft Research) released with the paper [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) by Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling.
 1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
 1. **[XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
 1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/main/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
+1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
+1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
 1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. Want to contribute a new model? We have added a **detailed guide and templates** to guide you in the process of adding a new model. You can find them in the [`templates`](./templates) folder of the repository. Be sure to check the [contributing guidelines](./CONTRIBUTING.md) and contact the maintainers or open an issue to collect feedbacks before starting your PR.
 

README_ko.md

@@ -221,16 +221,19 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
 1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
 1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
 1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
 1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
+1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
+1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
@@ -239,51 +242,67 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
 1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
 1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
+1. **[Donut](https://huggingface.co/docs/transformers/model_doc/donut)** (from NAVER) released with the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[ERNIE](https://huggingface.co/docs/transformers/model_doc/ernie)** (from Baidu) released with the paper [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/main/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
+1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
+1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
+1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
+1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
 1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
 1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
 1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
+1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
+1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
+1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
+1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[PEGASUS-X](https://huggingface.co/docs/transformers/model_doc/pegasus_x)** (from Google) released with the paper [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) by Jason Phang, Yao Zhao, Peter J. Liu. 
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
 1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
 1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
+1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](https://huggingface.co/docs/transformers/main/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
+1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
 1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
@@ -294,24 +313,30 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/main/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
 1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
+1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
 1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[VideoMAE](https://huggingface.co/docs/transformers/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
 1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
 1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
+1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
 1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
 1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[X-CLIP](https://huggingface.co/docs/transformers/model_doc/xclip)** (from Microsoft Research) released with the paper [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) by Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling.
 1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
@@ -320,7 +345,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
 1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/main/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
+1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
 1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. 새로운 모델을 올리고 싶나요? 우리가 **상세한 가이드와 템플릿** 으로 새로운 모델을 올리도록 도와드릴게요. 가이드와 템플릿은 이 저장소의 [`templates`](./templates) 폴더에서 확인하실 수 있습니다. [컨트리뷰션 가이드라인](./CONTRIBUTING.md)을 꼭 확인해주시고, PR을 올리기 전에 메인테이너에게 연락하거나 이슈를 오픈해 피드백을 받으시길 바랍니다. 
 

README_zh-hans.md

@@ -173,7 +173,7 @@ checkpoint: 检查点
     - 对所有模型统一的API
 
 1. 更低计算开销，更少的碳排放：
-    - 研究人员可以分享亿训练的模型而非次次从头开始训练
+    - 研究人员可以分享已训练的模型而非每次从头开始训练
     - 工程师可以减少计算用时和生产环境开销
     - 数十种模型架构、两千多个预训练模型、100多种语言支持
 
@@ -245,16 +245,19 @@ conda install -c huggingface transformers
 1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (来自 Google Research) 伴随论文 [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) 由 Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed 发布。
 1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (来自 Facebook) 伴随论文 [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) 由 Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston 发布。
 1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (来自 Facebook) 伴随论文 [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) 由 Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston 发布。
+1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/). 
 1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (来自 Alexa) 伴随论文 [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) 由 Adrian de Wynter and Daniel J. Perry 发布。
 1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (来自 Google Research) 伴随论文 [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) 由 Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel 发布。
 1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (来自 Inria/Facebook/Sorbonne) 伴随论文 [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) 由 Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot 发布。
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (来自 Google Research) 伴随论文 [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) 由 Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting 发布。
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (来自 OpenAI) 伴随论文 [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) 由 Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever 发布。
+1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (来自 Salesforce) 伴随论文 [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) 由 Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong 发布。
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (来自 YituTech) 伴随论文 [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) 由 Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan 发布。
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (来自 Facebook AI) 伴随论文 [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) 由 Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie 发布。
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (来自 Facebook AI) 伴随论文 [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) 由 Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie 发布。
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (来自 Tsinghua University) 伴随论文 [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) 由 Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun 发布。
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (来自 Salesforce) 伴随论文 [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) 由 Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher 发布。
-1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (来自 Facebook) 伴随论文 [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) 由 Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli 发布。
+1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (来自 Microsoft) 伴随论文 [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) 由 Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang 发布。
+1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (来自 Facebook) 伴随论文 [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) 由 Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli 发布。
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (来自 Microsoft) 伴随论文 [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) 由 Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen 发布。
 1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (来自 Berkeley/Facebook/Google) 伴随论文 [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) 由 Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch 发布。
@@ -263,51 +266,67 @@ conda install -c huggingface transformers
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (来自 Microsoft Research) 伴随论文 [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) 由 Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan 发布。
 1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (来自 HuggingFace), 伴随论文 [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) 由 Victor Sanh, Lysandre Debut and Thomas Wolf 发布。 同样的方法也应用于压缩 GPT-2 到 [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa 到 [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT 到 [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) 和德语版 DistilBERT。
 1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (来自 Microsoft Research) 伴随论文 [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) 由 Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei 发布。
+1. **[Donut](https://huggingface.co/docs/transformers/model_doc/donut)** (来自 NAVER) 伴随论文 [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) 由 Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park 发布。
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (来自 Facebook) 伴随论文 [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) 由 Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih 发布。
 1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (来自 Intel Labs) 伴随论文 [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) 由 René Ranftl, Alexey Bochkovskiy, Vladlen Koltun 发布。
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (来自 Google Research/Stanford University) 伴随论文 [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) 由 Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning 发布。
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (来自 Google Research) 伴随论文 [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) 由 Sascha Rothe, Shashi Narayan, Aliaksei Severyn 发布。
+1. **[ERNIE](https://huggingface.co/docs/transformers/model_doc/ernie)** (来自 Baidu) 伴随论文 [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu 发布。
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (来自 CNRS) 伴随论文 [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) 由 Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab 发布。
-1. **[FLAVA](https://huggingface.co/docs/transformers/main/model_doc/flava)** (来自 Facebook AI) 伴随论文 [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) 由 Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela 发布。
+1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (来自 Facebook AI) 伴随论文 [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) 由 Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela 发布。
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (来自 Google Research) 伴随论文 [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) 由 James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon 发布。
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (来自 CMU/Google Brain) 伴随论文 [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) 由 Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le 发布。
-1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (来自 KAIST) 伴随论文 [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) 由 Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim 发布。
+1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (来自 KAIST) 伴随论文 [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) 由 Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim 发布。
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (来自 OpenAI) 伴随论文 [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) 由 Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever 发布。
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (来自 EleutherAI) 随仓库 [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) 发布。作者为 Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy 发布。
+1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (来自 OpenAI) 伴随论文 [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) 由 Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever** 发布。
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (来自 EleutherAI) 伴随论文 [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) 由 Ben Wang and Aran Komatsuzaki 发布。
+1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (来自 UCSD, NVIDIA) 伴随论文 [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) 由 Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang 发布。
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (来自 Facebook) 伴随论文 [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) 由 Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed 发布。
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (来自 Berkeley) 伴随论文 [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) 由 Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer 发布。
-1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
+1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (来自 OpenAI) 伴随论文 [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) 由 Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever 发布。
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) 由 Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou 发布。
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) 由 Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou 发布。
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (来自 Microsoft Research Asia) 伴随论文 [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) 由 Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei 发布。
+1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (来自 Microsoft Research Asia) 伴随论文 [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) 由 Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei 发布。
+1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutxlm)** (来自 Microsoft Research Asia) 伴随论文 [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) 由 Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei 发布。
 1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (来自 AllenAI) 伴随论文 [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) 由 Iz Beltagy, Matthew E. Peters, Arman Cohan 发布。
+1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (来自 Meta AI) 伴随论文 [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) 由 Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze 发布。
 1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (来自 AllenAI) 伴随论文 [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) 由 Iz Beltagy, Matthew E. Peters, Arman Cohan 发布。
+1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (来自 Google AI) released 伴随论文 [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) 由 Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang 发布。
 1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (来自 Studio Ousia) 伴随论文 [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) 由 Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto 发布。
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (来自 UNC Chapel Hill) 伴随论文 [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) 由 Hao Tan and Mohit Bansal 发布。
+1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (来自 Facebook) 伴随论文 [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) 由 Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert 发布。
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (来自 Facebook) 伴随论文 [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) 由 Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin 发布。
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** 用 [OPUS](http://opus.nlpl.eu/) 数据训练的机器翻译模型由 Jörg Tiedemann 发布。[Marian Framework](https://marian-nmt.github.io/) 由微软翻译团队开发。
-1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
-1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) 由 Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer 发布。
-1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) 由 Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan 发布。
+1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
+1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) 由 Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer 发布。
+1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (来自 Facebook) 伴随论文 [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) 由 Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan 发布。
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (来自 Studio Ousia) 伴随论文 [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) 由 Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka 发布。
+1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (来自 CMU/Google Brain) 伴随论文 [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) 由 Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou 发布。
+1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (来自 Apple) 伴随论文 [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) 由 Sachin Mehta and Mohammad Rastegari 发布。
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (来自 Microsoft Research) 伴随论文 [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) 由 Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu 发布。
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (来自 Google AI) 伴随论文 [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) 由 Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel 发布。
+1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (来自 中国人民大学 AI Box) 伴随论文 [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) 由 Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen 发布。 
+1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (来自华为诺亚方舟实验室) 伴随论文 [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) 由 Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu 发布。
+1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (来自 Meta) 伴随论文 [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) 由 the NLLB team 发布。
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (来自 the University of Wisconsin - Madison) 伴随论文 [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) 由 Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh 发布。
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (来自 Meta AI) 伴随论文 [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) 由 Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al 发布。
+1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (来自 Google AI) 伴随论文 [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) 由 Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby 发布。
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (来自 Google) 伴随论文 [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) 由 Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu 发布。
+1. **[PEGASUS-X](https://huggingface.co/docs/transformers/model_doc/pegasus_x)** (来自 Google) 伴随论文 [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) 由 Jason Phang, Yao Zhao, Peter J. Liu 发布。  
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (来自 Deepmind) 伴随论文 [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) 由 Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira 发布。
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (来自 VinAI Research) 伴随论文 [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) 由 Dat Quoc Nguyen and Anh Tuan Nguyen 发布。
 1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (来自 UCLA NLP) 伴随论文 [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) 由 Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang 发布。
 1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (来自 Sea AI Labs) 伴随论文 [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) 由 Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng 发布。
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (来自 NVIDIA) 伴随论文 [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) 由 Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius 发布。
+1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (来自 Facebook) 伴随论文 [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) 由 Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela 发布。
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (来自 Google Research) 伴随论文 [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) 由 Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang 发布。
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (来自 Google Research) 伴随论文 [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) 由 Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya 发布。
-1. **[RegNet](https://huggingface.co/docs/transformers/main/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
+1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (来自 Google Research) 伴随论文 [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) 由 Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder 发布。
 1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (来自 Facebook), 伴随论文 [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) 由 Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov 发布。
@@ -318,24 +337,30 @@ conda install -c huggingface transformers
 1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (来自 Facebook), 伴随论文 [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) 由 Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino 发布。
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (来自 Facebook) 伴随论文 [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) 由 Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau 发布。
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (来自 Tel Aviv University) 伴随论文 [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) 由 Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy 发布。
-1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (来自 Berkeley) 伴随论文 [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) 由 Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer 发布。
+1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (来自 Berkeley) 伴随论文 [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) 由 Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer 发布。
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (来自 Microsoft) 伴随论文 [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) 由 Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo 发布。
+1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (来自 Microsoft) 伴随论文 [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) 由 Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo 发布。
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (来自 Google AI) 伴随论文 [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (来自 Google AI) 伴随论文 [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) 由 Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu 发布。
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (来自 Google AI) 伴随论文 [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) 由 Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos 发布。
-1. **[TAPEX](https://huggingface.co/docs/transformers/main/model_doc/tapex)** (来自 Microsoft Research) 伴随论文 [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) 由 Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou 发布。
+1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (来自 Microsoft Research) 伴随论文 [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) 由 Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou 发布。
+1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
 1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (来自 Google/CMU) 伴随论文 [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) 由 Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov 发布。
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (来自 Microsoft) 伴随论文 [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) 由 Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei 发布。
+1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (来自 Microsoft Research) 伴随论文 [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) 由 Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang 发布。
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (来自 Microsoft Research) 伴随论文 [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) 由 Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu 发布。
 1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (来自 Tsinghua University and Nankai University) 伴随论文 [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) 由 Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu 发布。
+1. **[VideoMAE](https://huggingface.co/docs/transformers/model_doc/videomae)** (来自 Multimedia Computing Group, Nanjing University) 伴随论文 [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) 由 Zhan Tong, Yibing Song, Jue Wang, Limin Wang 发布。
 1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (来自 NAVER AI Lab/Kakao Enterprise/Kakao Brain) 伴随论文 [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) 由 Wonjae Kim, Bokyung Son, Ildoo Kim 发布。
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (来自 Google AI) 伴随论文 [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) 由 Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby 发布。
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (来自 UCLA NLP) 伴随论文 [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) 由 Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang 发布。
 1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (来自 Meta AI) 伴随论文 [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) 由 Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick 发布。
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (来自 Facebook AI) 伴随论文 [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) 由 Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli 发布。
+1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (来自 Facebook AI) 伴随论文 [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) 由 Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino 发布。
 1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (来自 Facebook AI) 伴随论文 [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) 由 Qiantong Xu, Alexei Baevski, Michael Auli 发布。
 1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[X-CLIP](https://huggingface.co/docs/transformers/model_doc/xclip)** (来自 Microsoft Research) 伴随论文 [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) 由 Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling 发布。
 1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (来自 Facebook) 伴随论文 [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) 由 Guillaume Lample and Alexis Conneau 发布。
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (来自 Microsoft Research) 伴随论文 [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) 由 Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou 发布。
@@ -344,7 +369,7 @@ conda install -c huggingface transformers
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (来自 Google/CMU) 伴随论文 [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) 由 Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le 发布。
 1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (来自 Facebook AI) 伴随论文 [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) 由 Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli 发布。
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (来自 Facebook AI) 伴随论文 [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) 由 Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli 发布。
-1. **[YOLOS](https://huggingface.co/docs/transformers/main/model_doc/yolos)** (来自 Huazhong University of Science & Technology) 伴随论文 [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) 由 Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu 发布。
+1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (来自 Huazhong University of Science & Technology) 伴随论文 [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) 由 Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu 发布。
 1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (来自 the University of Wisconsin - Madison) 伴随论文 [You Only Sample (Almost) 由 Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh 发布。
 1. 想要贡献新的模型？我们这里有一份**详细指引和模板**来引导你添加新的模型。你可以在 [`templates`](./templates) 目录中找到他们。记得查看 [贡献指南](./CONTRIBUTING.md) 并在开始写 PR 前联系维护人员或开一个新的 issue 来获得反馈。
 
@@ -360,7 +385,7 @@ conda install -c huggingface transformers
 | [文档](https://huggingface.co/transformers/) | 完整的 API 文档和教程 |
 | [任务总结](https://huggingface.co/docs/transformers/task_summary) | 🤗 Transformers 支持的任务 |
 | [预处理教程](https://huggingface.co/docs/transformers/preprocessing) | 使用 `Tokenizer` 来为模型准备数据 |
-| [训练和微调](https://huggingface.co/docstransformers/training) | 在 PyTorch/TensorFlow 的训练循环或 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
+| [训练和微调](https://huggingface.co/docs/transformers/training) | 在 PyTorch/TensorFlow 的训练循环或 `Trainer` API 中使用 🤗 Transformers 提供的模型 |
 | [快速上手：微调和用例脚本](https://github.com/huggingface/transformers/tree/main/examples) | 为各种任务提供的用例脚本 |
 | [模型分享和上传](https://huggingface.co/docs/transformers/model_sharing) | 和社区上传和分享你微调的模型 |
 | [迁移](https://huggingface.co/docs/transformers/migration) | 从 `pytorch-transformers` 或 `pytorch-pretrained-bert` 迁移到 🤗 Transformers |

README_zh-hant.md

@@ -185,7 +185,7 @@ Tokenizer 為所有的預訓練模型提供了預處理，並可以直接轉換
     - 對所有模型使用的制式化API
 
 1. 更低的運算成本，更少的碳排放：
-    - 研究人員可以分享預訓練的模型而非從頭開始訓練
+    - 研究人員可以分享已訓練的模型而非每次從頭開始訓練
     - 工程師可以減少計算時間以及生產成本
     - 數十種模型架構、兩千多個預訓練模型、100多種語言支援
 
@@ -257,16 +257,19 @@ conda install -c huggingface transformers
 1. **[BigBird-RoBERTa](https://huggingface.co/docs/transformers/model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
 1. **[Blenderbot](https://huggingface.co/docs/transformers/model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
 1. **[BlenderbotSmall](https://huggingface.co/docs/transformers/model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BLOOM](https://huggingface.co/docs/transformers/model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/). 
 1. **[BORT](https://huggingface.co/docs/transformers/model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
 1. **[ByT5](https://huggingface.co/docs/transformers/model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](https://huggingface.co/docs/transformers/model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
 1. **[CLIP](https://huggingface.co/docs/transformers/model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
+1. **[CodeGen](https://huggingface.co/docs/transformers/model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[ConvBERT](https://huggingface.co/docs/transformers/model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
-1. **[ConvNeXT](https://huggingface.co/docs/transformers/main/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
+1. **[ConvNeXT](https://huggingface.co/docs/transformers/model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](https://huggingface.co/docs/transformers/model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](https://huggingface.co/docs/transformers/model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
-1. **[Data2Vec](https://huggingface.co/docs/transformers/main/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
+1. **[CvT](https://huggingface.co/docs/transformers/model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
+1. **[Data2Vec](https://huggingface.co/docs/transformers/model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](https://huggingface.co/docs/transformers/model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](https://huggingface.co/docs/transformers/model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[Decision Transformer](https://huggingface.co/docs/transformers/model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
@@ -275,51 +278,67 @@ conda install -c huggingface transformers
 1. **[DialoGPT](https://huggingface.co/docs/transformers/model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
 1. **[DistilBERT](https://huggingface.co/docs/transformers/model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/distillation) and a German version of DistilBERT.
 1. **[DiT](https://huggingface.co/docs/transformers/model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
+1. **[Donut](https://huggingface.co/docs/transformers/model_doc/donut)** (from NAVER) released with the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
 1. **[DPR](https://huggingface.co/docs/transformers/model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[DPT](https://huggingface.co/docs/transformers/master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
 1. **[ELECTRA](https://huggingface.co/docs/transformers/model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
 1. **[EncoderDecoder](https://huggingface.co/docs/transformers/model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[ERNIE](https://huggingface.co/docs/transformers/model_doc/ernie)** (from Baidu) released with the paper [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu.
 1. **[FlauBERT](https://huggingface.co/docs/transformers/model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
-1. **[FLAVA](https://huggingface.co/docs/transformers/main/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
+1. **[FLAVA](https://huggingface.co/docs/transformers/model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
 1. **[FNet](https://huggingface.co/docs/transformers/model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](https://huggingface.co/docs/transformers/model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
-1. **[GLPN](https://huggingface.co/docs/transformers/main/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
+1. **[GLPN](https://huggingface.co/docs/transformers/model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](https://huggingface.co/docs/transformers/model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
 1. **[GPT Neo](https://huggingface.co/docs/transformers/model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GPT NeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach 
 1. **[GPT-2](https://huggingface.co/docs/transformers/model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](https://huggingface.co/docs/transformers/model_doc/gptj)** (from EleutherAI) released with the paper [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
+1. **[GroupViT](https://huggingface.co/docs/transformers/model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[Hubert](https://huggingface.co/docs/transformers/model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](https://huggingface.co/docs/transformers/model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
-1. **[ImageGPT](https://huggingface.co/docs/transformers/main/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[ImageGPT](https://huggingface.co/docs/transformers/model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](https://huggingface.co/docs/transformers/model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
+1. **[LayoutLMv3](https://huggingface.co/docs/transformers/model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+1. **[LayoutXLM](https://huggingface.co/docs/transformers/model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
 1. **[LED](https://huggingface.co/docs/transformers/model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LeViT](https://huggingface.co/docs/transformers/model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
 1. **[Longformer](https://huggingface.co/docs/transformers/model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LongT5](https://huggingface.co/docs/transformers/model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
 1. **[LUKE](https://huggingface.co/docs/transformers/model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
 1. **[LXMERT](https://huggingface.co/docs/transformers/model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[M-CTC-T](https://huggingface.co/docs/transformers/model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
 1. **[M2M100](https://huggingface.co/docs/transformers/model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](https://huggingface.co/docs/transformers/model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
-1. **[MaskFormer](https://huggingface.co/docs/transformers/main/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
-1. **[MBart](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[MBart-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
+1. **[MaskFormer](https://huggingface.co/docs/transformers/model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov 
+1. **[mBART](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[mBART-50](https://huggingface.co/docs/transformers/model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](https://huggingface.co/docs/transformers/model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
+1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](https://huggingface.co/docs/transformers/model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[MVP](https://huggingface.co/docs/transformers/model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+1. **[Nezha](https://huggingface.co/docs/transformers/model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
+1. **[NLLB](https://huggingface.co/docs/transformers/model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
 1. **[Nyströmformer](https://huggingface.co/docs/transformers/model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[OPT](https://huggingface.co/docs/transformers/master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
+1. **[OWL-ViT](https://huggingface.co/docs/transformers/model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
 1. **[Pegasus](https://huggingface.co/docs/transformers/model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[PEGASUS-X](https://huggingface.co/docs/transformers/model_doc/pegasus_x)** (from Google) released with the paper [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) by Jason Phang, Yao Zhao, Peter J. Liu.
 1. **[Perceiver IO](https://huggingface.co/docs/transformers/model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](https://huggingface.co/docs/transformers/model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
 1. **[PLBart](https://huggingface.co/docs/transformers/model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
 1. **[PoolFormer](https://huggingface.co/docs/transformers/model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](https://huggingface.co/docs/transformers/model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](https://huggingface.co/docs/transformers/model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
+1. **[RAG](https://huggingface.co/docs/transformers/model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](https://huggingface.co/docs/transformers/model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](https://huggingface.co/docs/transformers/model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RegNet](https://huggingface.co/docs/transformers/main/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
+1. **[RegNet](https://huggingface.co/docs/transformers/model_doc/regnet)** (from META Research) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár. 
 1. **[RemBERT](https://huggingface.co/docs/transformers/model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/pdf/2010.12821.pdf) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
 1. **[ResNet](https://huggingface.co/docs/transformers/model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. 
 1. **[RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta)** (from Facebook), released together with the paper a [Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
@@ -330,24 +349,30 @@ conda install -c huggingface transformers
 1. **[SpeechToTextTransformer](https://huggingface.co/docs/transformers/model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
 1. **[SpeechToTextTransformer2](https://huggingface.co/docs/transformers/model_doc/speech_to_text_2)** (from Facebook) released with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](https://huggingface.co/docs/transformers/model_doc/splinter)** (from Tel Aviv University) released with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBert](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[SqueezeBERT](https://huggingface.co/docs/transformers/model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
 1. **[Swin Transformer](https://huggingface.co/docs/transformers/model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer V2](https://huggingface.co/docs/transformers/model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
 1. **[T5](https://huggingface.co/docs/transformers/model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](https://huggingface.co/docs/transformers/model_doc/t5v1.1)** (from Google AI) released with the paper [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](https://huggingface.co/docs/transformers/model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
-1. **[TAPEX](https://huggingface.co/docs/transformers/main/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[TAPEX](https://huggingface.co/docs/transformers/model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[Trajectory Transformer](https://huggingface.co/docs/transformers/model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine 
 1. **[Transformer-XL](https://huggingface.co/docs/transformers/model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
 1. **[TrOCR](https://huggingface.co/docs/transformers/model_doc/trocr)** (from Microsoft) released with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
+1. **[UL2](https://huggingface.co/docs/transformers/model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler 
 1. **[UniSpeech](https://huggingface.co/docs/transformers/model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](https://huggingface.co/docs/transformers/model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
 1. **[VAN](https://huggingface.co/docs/transformers/model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/pdf/2202.09741.pdf) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[VideoMAE](https://huggingface.co/docs/transformers/model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
 1. **[ViLT](https://huggingface.co/docs/transformers/model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](https://huggingface.co/docs/transformers/model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby. 
 1. **[VisualBERT](https://huggingface.co/docs/transformers/model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
 1. **[ViTMAE](https://huggingface.co/docs/transformers/model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
+1. **[Wav2Vec2-Conformer](https://huggingface.co/docs/transformers/model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
 1. **[Wav2Vec2Phoneme](https://huggingface.co/docs/transformers/model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
 1. **[WavLM](https://huggingface.co/docs/transformers/model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[X-CLIP](https://huggingface.co/docs/transformers/model_doc/xclip)** (from Microsoft Research) released with the paper [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) by Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling.
 1. **[XGLM](https://huggingface.co/docs/transformers/model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li. 
 1. **[XLM](https://huggingface.co/docs/transformers/model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](https://huggingface.co/docs/transformers/model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
@@ -356,7 +381,7 @@ conda install -c huggingface transformers
 1. **[XLNet](https://huggingface.co/docs/transformers/model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
 1. **[XLS-R](https://huggingface.co/docs/transformers/model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
 1. **[XLSR-Wav2Vec2](https://huggingface.co/docs/transformers/model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
-1. **[YOLOS](https://huggingface.co/docs/transformers/main/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
+1. **[YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
 1. **[YOSO](https://huggingface.co/docs/transformers/model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 1. 想要貢獻新的模型？我們這裡有一份**詳細指引和模板**來引導你加入新的模型。你可以在 [`templates`](./templates) 目錄中找到它們。記得查看[貢獻指引](./CONTRIBUTING.md)並在開始寫 PR 前聯繫維護人員或開一個新的 issue 來獲得 feedbacks。
 

docker/transformers-all-latest-gpu/Dockerfile

@@ -3,20 +3,53 @@ LABEL maintainer="Hugging Face"
 
 ARG DEBIAN_FRONTEND=noninteractive
 
+# Use login shell to read variables from `~/.profile` (to pass dynamic created variables between RUN commands)
+SHELL ["sh", "-lc"]
+
+# The following `ARG` are mainly used to specify the versions explicitly & directly in this docker file, and not meant
+# to be used as arguments for docker build (so far).
+
+ARG PYTORCH='1.12.1'
+# (not always a valid torch version)
+ARG INTEL_TORCH_EXT='1.11.0'
+# Example: `cu102`, `cu113`, etc.
+ARG CUDA='cu113'
+
 RUN apt update
-RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg
+RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg git-lfs
+RUN git lfs install
 RUN python3 -m pip install --no-cache-dir --upgrade pip
 
 ARG REF=main
 RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
 RUN python3 -m pip install --no-cache-dir -e ./transformers[dev,onnxruntime]
 
-RUN python3 -m pip install --no-cache-dir -U torch tensorflow
+# TODO: Handle these in a python utility script
+RUN [ ${#PYTORCH} -gt 0 -a "$PYTORCH" != "pre" ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; echo "export VERSION='$VERSION'" >> ~/.profile
+RUN echo torch=$VERSION
+# `torchvision` and `torchaudio` should be installed along with `torch`, especially for nightly build.
+# Currently, let's just use their latest releases (when `torch` is installed with a release version)
+# TODO: We might need to specify proper versions that work with a specific torch version (especially for past CI).
+RUN [ "$PYTORCH" != "pre" ] && python3 -m pip install --no-cache-dir -U $VERSION torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/$CUDA || python3 -m pip install --no-cache-dir -U --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/$CUDA
+
+RUN python3 -m pip install --no-cache-dir -U tensorflow
 RUN python3 -m pip uninstall -y flax jax
-RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu102.html
+
+# Use installed torch version for `torch-scatter` to avid to deal with PYTORCH='pre'.
+# If torch is nightly version, the link is likely to be invalid, but the installation falls back to the latest torch-scatter
+RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+$CUDA.html
+RUN python3 -m pip install --no-cache-dir intel_extension_for_pytorch==$INTEL_TORCH_EXT+cpu -f https://software.intel.com/ipex-whl-stable
+
 RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
 RUN python3 -m pip install -U "itsdangerous<2.1.0"
 
+RUN python3 -m pip install --no-cache-dir git+https://github.com/huggingface/accelerate@main#egg=accelerate
+
+# Add bitsandbytes for mixed int8 testing
+RUN python3 -m pip install --no-cache-dir bitsandbytes
+
+RUN python3 -m pip install --no-cache-dir decord
+
 # When installing in editable mode, `transformers` is not recognized as a package.
 # this line must be added in order for python to be aware of transformers.
 RUN cd transformers && python3 setup.py develop

docker/transformers-cpu/Dockerfile

@@ -23,4 +23,4 @@ COPY . transformers/
 RUN cd transformers/ && \
     python3 -m pip install --no-cache-dir .
 
-CMD ["/bin/bash"]
+CMD ["/bin/bash"]

docker/transformers-doc-builder/Dockerfile

@@ -15,5 +15,6 @@ RUN python3 -m pip install --no-cache-dir torchvision git+https://github.com/fac
 RUN python3 -m pip install --no-cache-dir pytorch-quantization --extra-index-url https://pypi.ngc.nvidia.com
 RUN python3 -m pip install -U "itsdangerous<2.1.0"
 
-RUN doc-builder build transformers transformers/docs/source --build_dir doc-build-dev --notebook_dir notebooks/transformers_doc --clean --version pr_$PR_NUMBER
+# Test if the image could successfully build the doc. before publishing the image
+RUN doc-builder build transformers transformers/docs/source/en --build_dir doc-build-dev --notebook_dir notebooks/transformers_doc --clean
 RUN rm -rf doc-build-dev

docker/transformers-past-gpu/Dockerfile

@@ -0,0 +1,43 @@
+ARG BASE_DOCKER_IMAGE="nvidia/cuda:11.2.2-cudnn8-devel-ubuntu20.04"
+FROM $BASE_DOCKER_IMAGE
+LABEL maintainer="Hugging Face"
+
+ARG DEBIAN_FRONTEND=noninteractive
+
+# Use login shell to read variables from `~/.profile` (to pass dynamic created variables between RUN commands)
+SHELL ["sh", "-lc"]
+
+RUN apt update
+RUN apt install -y git libsndfile1-dev tesseract-ocr espeak-ng python3 python3-pip ffmpeg git-lfs
+RUN git lfs install
+RUN python3 -m pip install --no-cache-dir --upgrade pip
+
+ARG REF=main
+RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
+RUN python3 -m pip install --no-cache-dir -e ./transformers[dev,onnxruntime]
+
+# When installing in editable mode, `transformers` is not recognized as a package.
+# this line must be added in order for python to be aware of transformers.
+RUN cd transformers && python3 setup.py develop
+
+ARG FRAMEWORK
+ARG VERSION
+
+# Remove all frameworks
+# (`accelerate` requires `torch`, and this causes import issues for TF-only testing)
+RUN python3 -m pip uninstall -y torch torchvision torchaudio accelerate tensorflow jax flax
+
+# Get the libraries and their versions to install, and write installation command to `~/.profile`.
+RUN python3 ./transformers/utils/past_ci_versions.py --framework $FRAMEWORK --version $VERSION
+
+# Install the target framework
+RUN echo "INSTALL_CMD = $INSTALL_CMD"
+RUN $INSTALL_CMD
+
+# Having installation problems for torch-scatter with torch <= 1.6. Disable so we have the same set of tests.
+# (This part will be removed once the logic of using `past_ci_versions.py` is used in other Dockerfile files.)
+# # Use installed torch version for `torch-scatter`.
+# # (The env. variable $CUDA is defined in `past_ci_versions.py`)
+# RUN [ "$FRAMEWORK" = "pytorch" ] && python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+$CUDA.html || echo "torch-scatter not to be installed"
+
+RUN python3 -m pip install -U "itsdangerous<2.1.0"

docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile

@@ -3,16 +3,30 @@ LABEL maintainer="Hugging Face"
 
 ARG DEBIAN_FRONTEND=noninteractive
 
+ARG PYTORCH='1.12.1'
+# Example: `cu102`, `cu113`, etc.
+ARG CUDA='cu113'
+
 RUN apt -y update
 RUN apt install -y libaio-dev
 RUN python3 -m pip install --no-cache-dir --upgrade pip
 
 ARG REF=main
 RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
-RUN python3 -m pip install --no-cache-dir -e ./transformers[deepspeed-testing]
 
-RUN git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build && \
-    DS_BUILD_CPU_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install -e . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
+# Install latest release PyTorch
+# (PyTorch must be installed before pre-compiling any DeepSpeed c++/cuda ops.)
+# (https://www.deepspeed.ai/tutorials/advanced-install/#pre-install-deepspeed-ops)
+RUN python3 -m pip install --no-cache-dir -U torch==$PYTORCH torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/$CUDA
+
+RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed-testing]
+
+# Pre-build **latest** DeepSpeed, so it would be ready for testing (otherwise, the 1st deepspeed test will timeout)
+RUN python3 -m pip uninstall -y deepspeed
+# This has to be run (again) inside the GPU VMs running the tests.
+# The installation works here, but some tests fail, if we don't pre-build deepspeed again in the VMs running the tests.
+# TODO: Find out why test fail.
+RUN DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install deepspeed --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
 
 # When installing in editable mode, `transformers` is not recognized as a package.
 # this line must be added in order for python to be aware of transformers.

docker/transformers-pytorch-deepspeed-nightly-gpu/Dockerfile

@@ -0,0 +1,35 @@
+FROM nvcr.io/nvidia/pytorch:21.03-py3
+LABEL maintainer="Hugging Face"
+
+ARG DEBIAN_FRONTEND=noninteractive
+
+# Example: `cu102`, `cu113`, etc.
+ARG CUDA='cu113'
+
+RUN apt -y update
+RUN apt install -y libaio-dev
+RUN python3 -m pip install --no-cache-dir --upgrade pip
+
+ARG REF=main
+RUN git clone https://github.com/huggingface/transformers && cd transformers && git checkout $REF
+
+# Install **nightly** release PyTorch (flag `--pre`)
+# (PyTorch must be installed before pre-compiling any DeepSpeed c++/cuda ops.)
+# (https://www.deepspeed.ai/tutorials/advanced-install/#pre-install-deepspeed-ops)
+RUN python3 -m pip install --no-cache-dir -U --pre torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/nightly/$CUDA
+
+RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed-testing]
+
+# Pre-build **nightly** release of DeepSpeed, so it would be ready for testing (otherwise, the 1st deepspeed test will timeout)
+RUN python3 -m pip uninstall -y deepspeed
+# This has to be run inside the GPU VMs running the tests. (So far, it fails here due to GPU checks during compilation.)
+# Issue: https://github.com/microsoft/DeepSpeed/issues/2010
+# RUN git clone https://github.com/microsoft/DeepSpeed && cd DeepSpeed && rm -rf build && \
+#    DS_BUILD_CPU_ADAM=1 DS_BUILD_FUSED_ADAM=1 DS_BUILD_AIO=1 DS_BUILD_UTILS=1 python3 -m pip install . --global-option="build_ext" --global-option="-j8" --no-cache -v --disable-pip-version-check 2>&1
+
+# When installing in editable mode, `transformers` is not recognized as a package.
+# this line must be added in order for python to be aware of transformers.
+RUN cd transformers && python3 setup.py develop
+
+# Disable for now as deepspeed is not installed above. To be enabled once the issue is fixed.
+# RUN python3 -c "from deepspeed.launcher.runner import main"

docker/transformers-pytorch-gpu/Dockerfile

@@ -12,12 +12,17 @@ RUN git clone https://github.com/huggingface/transformers && cd transformers &&
 RUN python3 -m pip install --no-cache-dir -e ./transformers[dev-torch,testing]
 
 # If set to nothing, will install the latest version
-ARG PYTORCH=''
+ARG PYTORCH='1.12.1'
+ARG TORCH_VISION=''
+ARG TORCH_AUDIO=''
+
+RUN [ ${#PYTORCH} -gt 0 ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
+RUN [ ${#TORCH_VISION} -gt 0 ] && VERSION='torchvision=='TORCH_VISION'.*' ||  VERSION='torchvision'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
+RUN [ ${#TORCH_AUDIO} -gt 0 ] && VERSION='torchaudio=='TORCH_AUDIO'.*' ||  VERSION='torchaudio'; python3 -m pip install --no-cache-dir -U $VERSION --extra-index-url https://download.pytorch.org/whl/cu113
 
-RUN [ ${#PYTORCH} -gt 0 ] && VERSION='torch=='$PYTORCH'.*' ||  VERSION='torch'; python3 -m pip install --no-cache-dir -U $VERSION
 RUN python3 -m pip uninstall -y tensorflow flax
 
-RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu102.html
+RUN python3 -m pip install --no-cache-dir torch-scatter -f https://data.pyg.org/whl/torch-$(python3 -c "from torch import version; print(version.__version__.split('+')[0])")+cu113.html
 RUN python3 -m pip install --no-cache-dir git+https://github.com/facebookresearch/detectron2.git pytesseract https://github.com/kpu/kenlm/archive/master.zip
 RUN python3 -m pip install -U "itsdangerous<2.1.0"
 

docker/transformers-tensorflow-cpu/Dockerfile

@@ -22,4 +22,4 @@ COPY . transformers/
 RUN cd transformers/ && \
     python3 -m pip install --no-cache-dir .
 
-CMD ["/bin/bash"]
+CMD ["/bin/bash"]

docs/README.md

@@ -43,18 +43,39 @@ Once you have setup the `doc-builder` and additional packages, you can generate
 typing the following command:
 
 ```bash
-doc-builder build transformers docs/source/ --build_dir ~/tmp/test-build
+doc-builder build transformers docs/source/en/ --build_dir ~/tmp/test-build
 ```
 
 You can adapt the `--build_dir` to set any temporary folder that you prefer. This command will create it and generate
 the MDX files that will be rendered as the documentation on the main website. You can inspect them in your favorite
 Markdown editor.
 
+## Previewing the documentation
+
+To preview the docs, first install the `watchdog` module with:
+
+```bash
+pip install watchdog
+```
+
+Then run the following command:
+
+```bash
+doc-builder preview {package_name} {path_to_docs}
+```
+
+For example:
+
+```bash
+doc-builder preview transformers docs/source/en/
+```
+
+The docs will be viewable at [http://localhost:3000](http://localhost:3000). You can also preview the docs once you have opened a PR. You will see a bot add a comment to a link where the documentation with your changes lives.
+
 ---
 **NOTE**
 
-It's not possible to see locally how the final documentation will look like for now. Once you have opened a PR, you
-will see a bot add a comment to a link where the documentation with your changes lives.
+The `preview` command only works with existing doc files. When you add a completely new file, you need to update `_toctree.yml` & restart `preview` command (`ctrl-c` to stop it & call `doc-builder preview ...` again).
 
 ---
 
@@ -407,4 +428,4 @@ Here are a few tips to help you debug the doctests and make them pass:
   * whitespace: one give whitespace (space, tabulation, new line) is equivalent to any number of whitespace, so you can add new lines where there are spaces to make your output more readable.
   * numerical values: you should never put more than 4 or 5 digits to expected results as different setups or library versions might get you slightly different results. `doctest` is configure to ignore any difference lower than the precision to which you wrote (so 1e-4 if you write 4 digits).
 - Don't leave a block of code that is very long to execute. If you can't make it fast, you can either not use the doctest syntax on it (so that it's ignored), or if you want to use the doctest syntax to show the results, you can add a comment `# doctest: +SKIP` at the end of the lines of code too long to execute
-- Each line of code that produces a result needs to have that result written below. You can ignore an output if you don't want to show it in your code example by adding a comment ` # doctest: +IGNORE_RESULT` at the end of the line of code produing it.
+- Each line of code that produces a result needs to have that result written below. You can ignore an output if you don't want to show it in your code example by adding a comment ` # doctest: +IGNORE_RESULT` at the end of the line of code producing it.

docs/TRANSLATING.md

@@ -54,5 +54,4 @@ The fields you should add are `local` (with the name of the file containing the
 
 Once you have translated the `_toctree.yml` file, you can start translating the [MDX](https://mdxjs.com/) files associated with your docs chapter.
 
-> 🙋 If you'd like others to help you with the translation, you can either [open an issue](https://github.com/huggingface/transformers/issues) or tag @[espejelomar](https://twitter.com/espejelomar)
- on Twitter to gain some visibility.
+> 🙋 If you'd like others to help you with the translation, you should [open an issue](https://github.com/huggingface/transformers/issues) and tag @sgugger.

docs/source/de/_config.py

@@ -0,0 +1,14 @@
+# docstyle-ignore
+INSTALL_CONTENT = """
+# Transformers installation
+! pip install transformers datasets
+# To install from source instead of the last release, comment the command above and uncomment the following one.
+# ! pip install git+https://github.com/huggingface/transformers.git
+"""
+
+notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
+black_avoid_patterns = {
+    "{processor_class}": "FakeProcessorClass",
+    "{model_class}": "FakeModelClass",
+    "{object_class}": "FakeObjectClass",
+}

docs/source/de/_toctree.yml

@@ -0,0 +1,12 @@
+- sections:
+  - local: index
+    title: 🤗 Transformers
+  - local: quicktour
+    title: Schnellstart
+  - local: installation
+    title: Installation
+  title: Erste Schritte
+- sections:
+  - local: pipeline_tutorial
+    title: Pipelines für Inferenzen
+  title: Tutorials

docs/source/de/index.mdx

@@ -0,0 +1,322 @@
+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# 🤗 Transformers
+
+Maschinelles Lernen auf dem neuesten Stand der Technik für PyTorch, TensorFlow und JAX.
+
+🤗 Transformers bietet APIs zum einfachen Herunterladen und Trainieren von vortrainierten Modellen auf dem neuesten Stand der Technik. Die Verwendung von vortrainierten Modellen kann Rechenkosten sparen und den CO2-Fußabdruck reduzieren und Zeit sparen, die für das Training eines Modells von Grund auf benötigt wird. Die Modelle können für verschiedene Modalitäten verwendet werden, wie z. B.:
+
+* 📝 Text: Textklassifizierung, Informationsextrahierung, Beantwortung von Fragen, Zusammenfassung, Übersetzung und Texterstellung in über 100 Sprachen.
+* 🖼️ Bilder: Bildklassifizierung, Objekterkennung und Segmentierung.
+* 🗣️ Audio: Spracherkennung und Audioklassifizierung.
+* 🐙 Multimodal: Beantwortung von Tabellenfragen, optische Zeichenerkennung, Informationsextraktion aus gescannten Dokumenten, Videoklassifizierung und Beantwortung visueller Fragen.
+
+Unsere Bibliothek unterstützt die nahtlose Integration von drei der beliebtesten Deep-Learning-Bibliotheken: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) und [JAX](https://jax.readthedocs.io/en/latest/). Trainieren Sie Ihr Modell in drei Codezeilen in einem Framework und laden Sie es zur Inferenz mit einem anderen.
+
+Jede 🤗 Transformers-Architektur ist in einem eigenständigen Python-Modul definiert, so dass sie leicht für Forschung und Experimente angepasst werden kann.
+
+## Wenn Sie auf der Suche nach individueller Unterstützung durch das Hugging Face-Team sind
+
+<a target="_blank" href="https://huggingface.co/support">
+    <img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
+</a><br>
+
+## Inhalt
+
+Die Dokumentation ist in fünf Teile gegliedert:
+
+- **GET STARTED** enthält eine kurze Tour und Installationsanweisungen, um mit 🤗 Transformers loszulegen.
+- **TUTORIALS** sind ein hervorragender Ausgangspunkt, wenn Sie neu in unserer Bibliothek sind. Dieser Abschnitt hilft Ihnen, die grundlegenden Fähigkeiten zu erlangen, die Sie benötigen, um mit 🤗 Transformers zu arbeiten.
+- **HOW-TO GUIDES** zeigen Ihnen, wie Sie ein bestimmtes Ziel erreichen können, z. B. die Feinabstimmung eines vortrainierten Modells für die Sprachmodellierung oder die Erstellung eines benutzerdefinierten Modellkopfs.
+- **KONZEPTUELLE ANLEITUNGEN** bietet weitere Diskussionen und Erklärungen zu den zugrunde liegenden Konzepten und Ideen hinter Modellen, Aufgaben und der Designphilosophie von 🤗 Transformers. 
+- **API** beschreibt jede Klasse und Funktion, gruppiert in:
+
+  - **MAIN CLASSES** für die Hauptklassen, die die wichtigsten APIs der Bibliothek darstellen.
+  - MODELLE** für die Klassen und Funktionen, die zu jedem in der Bibliothek implementierten Modell gehören.
+  - **INTERNAL HELPERS** für die Klassen und Funktionen, die wir intern verwenden.
+
+Die Bibliothek enthält derzeit JAX-, PyTorch- und TensorFlow-Implementierungen, vortrainierte Modellgewichte, Nutzungsskripte und Konvertierungsprogramme für die folgenden Modelle.
+
+### Unterstütze Modelle
+
+<!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
+
+1. **[ALBERT](model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
+1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
+1. **[BARThez](model_doc/barthez)** (from École polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
+1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
+1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
+1. **[BERT](model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
+1. **[BERT For Sequence Generation](model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
+1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[Blenderbot](model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BlenderbotSmall](model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BLOOM](model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
+1. **[BORT](model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
+1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
+1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
+1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
+1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
+1. **[CodeGen](model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
+1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
+1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
+1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
+1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[CvT](model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
+1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
+1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
+1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
+1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
+1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
+1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
+1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
+1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
+1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
+1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
+1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
+1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
+1. **[FLAVA](model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
+1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
+1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
+1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
+1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
+1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GPT NeoX](model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
+1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
+1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
+1. **[GroupViT](model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
+1. **[Hubert](model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
+1. **[I-BERT](model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
+1. **[ImageGPT](model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
+1. **[LayoutLM](model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
+1. **[LayoutLMv2](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
+1. **[LayoutLMv3](model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+1. **[LayoutXLM](model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
+1. **[LED](model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LeViT](model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
+1. **[Longformer](model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LongT5](model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
+1. **[LUKE](model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
+1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[M-CTC-T](model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
+1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
+1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
+1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
+1. **[mBART](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[mBART-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
+1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
+1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
+1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
+1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
+1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
+1. **[MT5](model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[MVP](model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+1. **[Nezha](model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
+1. **[NLLB](model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
+1. **[Nyströmformer](model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
+1. **[OPT](master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
+1. **[OWL-ViT](model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
+1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
+1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
+1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
+1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
+1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
+1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
+1. **[RAG](model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
+1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
+1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
+1. **[RegNet](model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
+1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
+1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
+1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
+1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
+1. **[SegFormer](model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
+1. **[SEW](model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
+1. **[SEW-D](model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
+1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
+1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
+1. **[Splinter](model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
+1. **[SqueezeBERT](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[Swin Transformer](model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer V2](model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
+1. **[T5](model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
+1. **[T5v1.1](model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
+1. **[TAPAS](model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
+1. **[TAPEX](model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
+1. **[Transformer-XL](model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
+1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
+1. **[UL2](model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
+1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
+1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
+1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[VideoMAE](model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
+1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
+1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
+1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
+1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
+1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
+1. **[Wav2Vec2-Conformer](model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
+1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
+1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
+1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
+1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
+1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
+1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
+1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
+1. **[YOLOS](model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
+1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
+
+
+### Unterstützte Frameworks
+
+Die folgende Tabelle zeigt die derzeitige Unterstützung in der Bibliothek für jedes dieser Modelle, unabhängig davon, ob sie einen Python
+Tokenizer haben (als "langsam" bezeichnet), ein "schneller" Tokenizer, der von der 🤗 Tokenizers Bibliothek unterstützt wird, ob sie Unterstützung in Jax (via
+Flax), PyTorch, und/oder TensorFlow haben.
+
+<!--This table is updated automatically from the auto modules with _make fix-copies_. Do not update manually!-->
+
+|            Model            | Tokenizer slow | Tokenizer fast | PyTorch support | TensorFlow support | Flax Support |
+|:---------------------------:|:--------------:|:--------------:|:---------------:|:------------------:|:------------:|
+|           ALBERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|            BART             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|            BEiT             |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|            BERT             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|       Bert Generation       |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           BigBird           |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
+|       BigBird-Pegasus       |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         Blenderbot          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|       BlenderbotSmall       |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|            BLOOM            |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
+|          CamemBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           CANINE            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            CLIP             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|           CodeGen           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|          ConvBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|          ConvNeXT           |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|            CTRL             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|             CvT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Data2VecAudio        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Data2VecText         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|       Data2VecVision        |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|           DeBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|         DeBERTa-v2          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|    Decision Transformer     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            DeiT             |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|            DETR             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         DistilBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|             DPR             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|             DPT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           ELECTRA           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|       Encoder decoder       |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+| FairSeq Machine-Translation |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          FlauBERT           |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|            FLAVA            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            FNet             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|     Funnel Transformer      |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|            GLPN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           GPT Neo           |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|          GPT NeoX           |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
+|            GPT-J            |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|          GroupViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           Hubert            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|           I-BERT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          ImageGPT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          LayoutLM           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|         LayoutLMv2          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|         LayoutLMv3          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|             LED             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|            LeViT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         Longformer          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           LongT5            |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|            LUKE             |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           LXMERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           M-CTC-T           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           M2M100            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           Marian            |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
+|         MaskFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            mBART            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|        Megatron-BERT        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         MobileBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|          MobileViT          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            MPNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|             MT5             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|             MVP             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|            Nezha            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Nyströmformer        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         OpenAI GPT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|        OpenAI GPT-2         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|             OPT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|           OWL-ViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           Pegasus           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|          Perceiver          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           PLBart            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         PoolFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         ProphetNet          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           QDQBert           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             RAG             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|            REALM            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|          Reformer           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|           RegNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|           RemBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           ResNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|          RetriBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|           RoBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|          RoFormer           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|          SegFormer          |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|             SEW             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            SEW-D            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|   Speech Encoder decoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|         Speech2Text         |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|        Speech2Text2         |       ✅       |       ❌       |       ❌        |         ❌         |      ❌      |
+|          Splinter           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|         SqueezeBERT         |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|      Swin Transformer       |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|     Swin Transformer V2     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             T5              |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|            TAPAS            |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|   Trajectory Transformer    |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|       Transformer-XL        |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|            TrOCR            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          UniSpeech          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        UniSpeechSat         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             VAN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          VideoMAE           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            ViLT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|   Vision Encoder decoder    |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|    VisionTextDualEncoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|         VisualBERT          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             ViT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|           ViTMAE            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|          Wav2Vec2           |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
+|     Wav2Vec2-Conformer      |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            WavLM            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            XGLM             |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
+|             XLM             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|       XLM-ProphetNet        |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         XLM-RoBERTa         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|       XLM-RoBERTa-XL        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            XLNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|            YOLOS            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            YOSO             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+
+<!-- End table-->

docs/source/de/installation.mdx

@@ -0,0 +1,246 @@
+<!---
+Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+-->
+
+# Installation
+
+Installieren Sie 🤗 Transformers für die Deep-Learning-Bibliothek, mit der Sie arbeiten, richten Sie Ihren Cache ein und konfigurieren Sie 🤗 Transformers optional für den Offline-Betrieb.
+
+🤗 Transformers wurde unter Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, und Flax getestet. Folgen Sie den Installationsanweisungen unten für die von Ihnen verwendete Deep-Learning-Bibliothek:
+
+* [PyTorch](https://pytorch.org/get-started/locally/) installation instructions.
+* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) installation instructions.
+* [Flax](https://flax.readthedocs.io/en/latest/) installation instructions.
+
+## Installation mit pip
+
+Sie sollten 🤗 Transformers in einer [virtuellen Umgebung](https://docs.python.org/3/library/venv.html) installieren. Wenn Sie mit virtuellen Python-Umgebungen nicht vertraut sind, werfen Sie einen Blick auf diese [Anleitung](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). Eine virtuelle Umgebung macht es einfacher, verschiedene Projekte zu verwalten und Kompatibilitätsprobleme zwischen Abhängigkeiten zu vermeiden.
+
+Beginnen wir mit der Erstellung einer virtuellen Umgebung in Ihrem Projektverzeichnis:
+
+
+```bash
+python -m venv .env
+```
+
+Aktivieren wir die virtuelle Umgebung. Unter Linux und MacOs:
+
+```bash
+source .env/bin/activate
+```
+Aktivieren wir die virtuelle Umgebung unter Windows
+
+```bash
+.env/Scripts/activate
+```
+
+Jetzt können wir die 🤗 Transformers mit dem folgenden Befehl installieren:
+
+```bash
+pip install transformers
+```
+
+Bei reiner CPU-Unterstützung können wir 🤗 Transformers und eine Deep-Learning-Bibliothek bequem in einer Zeile installieren. Installieren wir zum Beispiel 🤗 Transformers und PyTorch mit:
+
+```bash
+pip install transformers[torch]
+```
+
+🤗 Transformers und TensorFlow 2.0:
+
+```bash
+pip install transformers[tf-cpu]
+```
+
+🤗 Transformers und Flax:
+
+```bash
+pip install transformers[flax]
+```
+
+Überprüfen wir abschließend, ob 🤗 Transformers ordnungsgemäß installiert wurde, indem wir den folgenden Befehl ausführen. Es wird ein vortrainiertes Modell heruntergeladen:
+
+```bash
+python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
+```
+
+Dann wird die Kategorie und die Wahrscheinlichkeit ausgegeben:
+
+```bash
+[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
+```
+
+## Installation aus dem Code
+
+Installieren wir 🤗 Transformers aus dem Quellcode mit dem folgenden Befehl:
+
+```bash
+pip install git+https://github.com/huggingface/transformers
+```
+
+Dieser Befehl installiert die aktuelle `main` Version und nicht die neueste `stable` Version. Die `main`-Version ist nützlich, um mit den neuesten Entwicklungen Schritt zu halten. Zum Beispiel, wenn ein Fehler seit der letzten offiziellen Version behoben wurde, aber eine neue Version noch nicht veröffentlicht wurde. Das bedeutet jedoch, dass die "Hauptversion" nicht immer stabil ist. Wir bemühen uns, die Hauptversion einsatzbereit zu halten, und die meisten Probleme werden normalerweise innerhalb weniger Stunden oder eines Tages behoben. Wenn Sie auf ein Problem stoßen, öffnen Sie bitte ein [Issue] (https://github.com/huggingface/transformers/issues), damit wir es noch schneller beheben können!
+
+Überprüfen wir, ob 🤗 Transformers richtig installiert wurde, indem Sie den folgenden Befehl ausführen:
+
+
+```bash
+python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
+```
+
+## Editierbare Installation
+
+Sie benötigen eine bearbeitbare Installation, wenn Sie:
+
+* die "Haupt"-Version des Quellcodes verwenden möchten.
+* Zu 🤗 Transformers beitragen und Änderungen am Code testen wollen.
+
+Klonen Sie das Repository und installieren 🤗 Transformers mit den folgenden Befehlen:
+
+```bash
+git clone https://github.com/huggingface/transformers.git
+cd transformers
+pip install -e .
+```
+
+Diese Befehle verknüpfen den Ordner, in den Sie das Repository geklont haben, mit den Pfaden Ihrer Python-Bibliotheken. Python wird nun in dem Ordner suchen, in den Sie geklont haben, zusätzlich zu den normalen Bibliothekspfaden. Wenn zum Beispiel Ihre Python-Pakete normalerweise in `~/anaconda3/envs/main/lib/python3.7/site-packages/` installiert sind, wird Python auch den Ordner durchsuchen, in den Sie geklont haben: `~/transformers/`.
+
+
+<Tip warning={true}>
+
+Sie müssen den Ordner `transformers` behalten, wenn Sie die Bibliothek weiter verwenden wollen.
+
+</Tip>
+
+Jetzt können Sie Ihren Klon mit dem folgenden Befehl ganz einfach auf die neueste Version von 🤗 Transformers aktualisieren:
+
+
+```bash
+cd ~/transformers/
+git pull
+```
+
+Ihre Python-Umgebung wird beim nächsten Ausführen die `main`-Version von 🤗 Transformers finden.
+
+## Installation mit conda
+
+Installation von dem conda Kanal `huggingface`:
+
+```bash
+conda install -c huggingface transformers
+```
+
+## Cache Einrichtung
+
+Vorgefertigte Modelle werden heruntergeladen und lokal zwischengespeichert unter: `~/.cache/huggingface/hub`. Dies ist das Standardverzeichnis, das durch die Shell-Umgebungsvariable "TRANSFORMERS_CACHE" vorgegeben ist. Unter Windows wird das Standardverzeichnis durch `C:\Benutzer\Benutzername\.cache\huggingface\hub` angegeben. Sie können die unten aufgeführten Shell-Umgebungsvariablen - in der Reihenfolge ihrer Priorität - ändern, um ein anderes Cache-Verzeichnis anzugeben:
+
+1. Shell-Umgebungsvariable (Standard): `HUGGINGFACE_HUB_CACHE` oder `TRANSFORMERS_CACHE`.
+2. Shell-Umgebungsvariable: `HF_HOME`.
+3. Shell-Umgebungsvariable: `XDG_CACHE_HOME` + `/huggingface`.
+
+
+<Tip>
+
+Transformers verwendet die Shell-Umgebungsvariablen `PYTORCH_TRANSFORMERS_CACHE` oder `PYTORCH_PRETRAINED_BERT_CACHE`, wenn Sie von einer früheren Iteration dieser Bibliothek kommen und diese Umgebungsvariablen gesetzt haben, sofern Sie nicht die Shell-Umgebungsvariable `TRANSFORMERS_CACHE` angeben.
+  
+</Tip>
+
+## Offline Modus
+
+Transformers ist in der Lage, in einer Firewall- oder Offline-Umgebung zu laufen, indem es nur lokale Dateien verwendet. Setzen Sie die Umgebungsvariable `TRANSFORMERS_OFFLINE=1`, um dieses Verhalten zu aktivieren.
+
+<Tip>
+
+Fügen sie [🤗 Datasets](https://huggingface.co/docs/datasets/) zu Ihrem Offline-Trainingsworkflow hinzufügen, indem Sie die Umgebungsvariable `HF_DATASETS_OFFLINE=1` setzen.
+
+</Tip>
+
+So würden Sie beispielsweise ein Programm in einem normalen Netzwerk mit einer Firewall für externe Instanzen mit dem folgenden Befehl ausführen:
+
+```bash
+python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
+```
+
+Führen Sie das gleiche Programm in einer Offline-Instanz mit aus:
+
+```bash
+HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
+python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
+```
+
+Das Skript sollte nun laufen, ohne sich aufzuhängen oder eine Zeitüberschreitung abzuwarten, da es weiß, dass es nur nach lokalen Dateien suchen soll.
+
+
+### Abrufen von Modellen und Tokenizern zur Offline-Verwendung
+
+Eine andere Möglichkeit, 🤗 Transformers offline zu verwenden, besteht darin, die Dateien im Voraus herunterzuladen und dann auf ihren lokalen Pfad zu verweisen, wenn Sie sie offline verwenden müssen. Es gibt drei Möglichkeiten, dies zu tun:
+
+* Laden Sie eine Datei über die Benutzeroberfläche des [Model Hub](https://huggingface.co/models) herunter, indem Sie auf das ↓-Symbol klicken.
+
+    ![download-icon](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/download-icon.png)
+
+* Verwenden Sie den [PreTrainedModel.from_pretrained] und [PreTrainedModel.save_pretrained] Workflow:
+
+    1. Laden Sie Ihre Dateien im Voraus mit [`PreTrainedModel.from_pretrained`] herunter:
+
+    ```py
+    >>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+
+    >>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
+    >>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
+    ```
+
+    2. Speichern Sie Ihre Dateien in einem bestimmten Verzeichnis mit [`PreTrainedModel.save_pretrained`]:
+
+    ```py
+    >>> tokenizer.save_pretrained("./your/path/bigscience_t0")
+    >>> model.save_pretrained("./your/path/bigscience_t0")
+    ```
+
+    3. Wenn Sie nun offline sind, laden Sie Ihre Dateien mit [`PreTrainedModel.from_pretrained`] aus dem bestimmten Verzeichnis:
+
+    ```py
+    >>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
+    >>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
+    ```
+
+* Programmatisches Herunterladen von Dateien mit der [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) Bibliothek:
+
+    1. Installieren Sie die "huggingface_hub"-Bibliothek in Ihrer virtuellen Umgebung:
+
+    ```bash
+    python -m pip install huggingface_hub
+    ```
+
+    2. Verwenden Sie die Funktion [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub), um eine Datei in einen bestimmten Pfad herunterzuladen. Der folgende Befehl lädt zum Beispiel die Datei "config.json" aus dem Modell [T0](https://huggingface.co/bigscience/T0_3B) in den gewünschten Pfad herunter:
+
+    ```py
+    >>> from huggingface_hub import hf_hub_download
+
+    >>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
+    ```
+
+Sobald Ihre Datei heruntergeladen und lokal zwischengespeichert ist, geben Sie den lokalen Pfad an, um sie zu laden und zu verwenden:
+
+```py
+>>> from transformers import AutoConfig
+
+>>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
+```
+
+<Tip>
+
+Weitere Informationen zum Herunterladen von Dateien, die auf dem Hub gespeichert sind, finden Sie im Abschnitt [Wie man Dateien vom Hub herunterlädt] (https://huggingface.co/docs/hub/how-to-downstream).
+  
+</Tip>

docs/source/de/pipeline_tutorial.mdx

@@ -0,0 +1,171 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# Pipelines für Inferenzen
+
+Die [`pipeline`] macht es einfach, jedes beliebige Modell aus dem [Hub](https://huggingface.co/models) für die Inferenz auf jede Sprache, Computer Vision, Sprache und multimodale Aufgaben zu verwenden. Selbst wenn Sie keine Erfahrung mit einer bestimmten Modalität haben oder nicht mit dem zugrundeliegenden Code hinter den Modellen vertraut sind, können Sie sie mit der [`pipeline`] für Inferenzen verwenden! In diesem Beispiel lernen Sie, wie:
+
+* Eine [`pipeline`] für Inferenz zu verwenden.
+* Einen bestimmten Tokenizer oder ein bestimmtes Modell zu verwenden.
+* Eine [`pipeline`] für Audio-, Vision- und multimodale Aufgaben zu verwenden.
+
+<Tip>
+
+Eine vollständige Liste der unterstützten Aufgaben und verfügbaren Parameter finden Sie in der [`pipeline`]-Dokumentation.
+
+</Tip>
+
+## Verwendung von Pipelines
+
+Obwohl jede Aufgabe eine zugehörige [`pipeline`] hat, ist es einfacher, die allgemeine [`pipeline`]-Abstraktion zu verwenden, die alle aufgabenspezifischen Pipelines enthält. Die [`pipeline`] lädt automatisch ein Standardmodell und eine Vorverarbeitungsklasse, die für Ihre Aufgabe inferenzfähig ist.
+
+1. Beginnen Sie mit der Erstellung einer [`pipeline`] und geben Sie eine Inferenzaufgabe an:
+
+```py
+>>> from transformers import pipeline
+
+>>> generator = pipeline(task="text-generation")
+```
+
+2. Übergeben Sie Ihren Eingabetext an die [`pipeline`]:
+
+```py
+>>> generator(
+...     "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
+... )  # doctest: +SKIP
+[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Iron-priests at the door to the east, and thirteen for the Lord Kings at the end of the mountain'}]
+```
+
+Wenn Sie mehr als eine Eingabe haben, übergeben Sie die Eingabe als Liste:
+
+```py
+>>> generator(
+...     [
+...         "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
+...         "Nine for Mortal Men, doomed to die, One for the Dark Lord on his dark throne",
+...     ]
+... )  # doctest: +SKIP
+```
+
+Alle zusätzlichen Parameter für Ihre Aufgabe können auch in die [`pipeline`] aufgenommen werden. Die Aufgabe `Text-Generierung` hat eine [`~generation_utils.GenerationMixin.generate`]-Methode mit mehreren Parametern zur Steuerung der Ausgabe. Wenn Sie zum Beispiel mehr als eine Ausgabe erzeugen wollen, setzen Sie den Parameter `num_return_sequences`:
+
+```py
+>>> generator(
+...     "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
+...     num_return_sequences=2,
+... )  # doctest: +SKIP
+```
+
+### Wählen Sie ein Modell und einen Tokenizer
+
+Die [`pipeline`] akzeptiert jedes Modell aus dem [Hub] (https://huggingface.co/models). Auf dem Hub gibt es Tags, mit denen Sie nach einem Modell filtern können, das Sie für Ihre Aufgabe verwenden möchten. Sobald Sie ein passendes Modell ausgewählt haben, laden Sie es mit der entsprechenden `AutoModelFor` und [`AutoTokenizer`] Klasse. Laden Sie zum Beispiel die Klasse [`AutoModelForCausalLM`] für eine kausale Sprachmodellierungsaufgabe:
+
+```py
+>>> from transformers import AutoTokenizer, AutoModelForCausalLM
+
+>>> tokenizer = AutoTokenizer.from_pretrained("distilgpt2")
+>>> model = AutoModelForCausalLM.from_pretrained("distilgpt2")
+```
+
+Erstellen Sie eine [`pipeline`] für Ihre Aufgabe, und geben Sie das Modell und den Tokenizer an, die Sie geladen haben:
+
+```py
+>>> from transformers import pipeline
+
+>>> generator = pipeline(task="text-generation", model=model, tokenizer=tokenizer)
+```
+
+Übergeben Sie Ihren Eingabetext an die [`pipeline`] , um einen Text zu erzeugen:
+
+```py
+>>> generator(
+...     "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
+... )  # doctest: +SKIP
+[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Dragon-lords (for them to rule in a world ruled by their rulers, and all who live within the realm'}]
+```
+
+## Audio-Pipeline
+
+Die [`pipeline`] unterstützt auch Audioaufgaben wie Audioklassifizierung und automatische Spracherkennung.
+
+Lassen Sie uns zum Beispiel die Emotion in diesem Audioclip klassifizieren:
+
+```py
+>>> from datasets import load_dataset
+>>> import torch
+
+>>> torch.manual_seed(42)  # doctest: +IGNORE_RESULT
+>>> ds = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
+>>> audio_file = ds[0]["audio"]["path"]
+```
+
+Finden Sie ein [Audioklassifikation](https://huggingface.co/models?pipeline_tag=audio-classification) Modell auf dem Model Hub für Emotionserkennung und laden Sie es in die [`pipeline`]:
+
+```py
+>>> from transformers import pipeline
+
+>>> audio_classifier = pipeline(
+...     task="audio-classification", model="ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
+... )
+```
+
+Übergeben Sie die Audiodatei an die [`pipeline`]:
+
+```py
+>>> preds = audio_classifier(audio_file)
+>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
+>>> preds
+[{'score': 0.1315, 'label': 'calm'}, {'score': 0.1307, 'label': 'neutral'}, {'score': 0.1274, 'label': 'sad'}, {'score': 0.1261, 'label': 'fearful'}, {'score': 0.1242, 'label': 'happy'}]
+```
+
+## Bildverarbeitungs-Pipeline
+
+Die Verwendung einer [`pipeline`] für Bildverarbeitungsaufgaben ist praktisch identisch.
+
+Geben Sie Ihre Aufgabe an und übergeben Sie Ihr Bild an den Klassifikator. Das Bild kann ein Link oder ein lokaler Pfad zu dem Bild sein. Zum Beispiel: Welche Katzenart ist unten abgebildet?
+
+![pipeline-cat-chonk](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg)
+
+```py
+>>> from transformers import pipeline
+
+>>> vision_classifier = pipeline(task="image-classification")
+>>> preds = vision_classifier(
+...     images="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
+... )
+>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
+>>> preds
+[{'score': 0.4335, 'label': 'lynx, catamount'}, {'score': 0.0348, 'label': 'cougar, puma, catamount, mountain lion, painter, panther, Felis concolor'}, {'score': 0.0324, 'label': 'snow leopard, ounce, Panthera uncia'}, {'score': 0.0239, 'label': 'Egyptian cat'}, {'score': 0.0229, 'label': 'tiger cat'}]
+```
+
+## Multimodale Pipeline
+
+Die [`pipeline`] unterstützt mehr als eine Modalität. Eine Aufgabe zur Beantwortung visueller Fragen (VQA) kombiniert zum Beispiel Text und Bild. Verwenden Sie einen beliebigen Bildlink und eine Frage, die Sie zu dem Bild stellen möchten. Das Bild kann eine URL oder ein lokaler Pfad zu dem Bild sein.
+
+Wenn Sie zum Beispiel das gleiche Bild wie in der obigen Vision-Pipeline verwenden:
+
+```py
+>>> image = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
+>>> question = "Where is the cat?"
+```
+
+Erstellen Sie eine Pipeline für "vqa" und übergeben Sie ihr das Bild und die Frage:
+
+```py
+>>> from transformers import pipeline
+
+>>> vqa = pipeline(task="vqa")
+>>> preds = vqa(image=image, question=question)
+>>> preds = [{"score": round(pred["score"], 4), "answer": pred["answer"]} for pred in preds]
+>>> preds
+[{'score': 0.9112, 'answer': 'snow'}, {'score': 0.8796, 'answer': 'in snow'}, {'score': 0.6717, 'answer': 'outside'}, {'score': 0.0291, 'answer': 'on ground'}, {'score': 0.027, 'answer': 'ground'}]
+```

docs/source/de/quicktour.mdx

@@ -0,0 +1,428 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# Schnellstart
+
+[[open-in-colab]]
+
+Mit 🤗 Transformers können Sie sofort loslegen! Verwenden Sie die [`pipeline`] für schnelle Inferenz und laden Sie schnell ein vortrainiertes Modell und einen Tokenizer mit einer [AutoClass](./model_doc/auto), um Ihre Text-, Bild- oder Audioaufgabe zu lösen.
+
+<Tip>
+
+Alle in der Dokumentation vorgestellten Codebeispiele haben oben links einen Umschalter für PyTorch und TensorFlow. Wenn
+nicht, wird erwartet, dass der Code für beide Backends ohne Änderungen funktioniert.
+
+</Tip>
+
+## Pipeline
+
+[`pipeline`] ist der einfachste Weg, ein vortrainiertes Modell für eine bestimmte Aufgabe zu verwenden.
+
+<Youtube id="tiZFewofSLM"/>
+
+Die [`pipeline`] unterstützt viele gängige Aufgaben:
+
+**Text**:
+* Stimmungsanalyse: Klassifizierung der Polarität eines gegebenen Textes.
+* Textgenerierung (auf Englisch): Generierung von Text aus einer gegebenen Eingabe.
+* Name-Entity-Recognition (NER): Kennzeichnung jedes Worts mit der Entität, die es repräsentiert (Person, Datum, Ort usw.).
+* Beantwortung von Fragen: Extrahieren der Antwort aus dem Kontext, wenn ein gewisser Kontext und eine Frage gegeben sind.
+* Fill-mask: Ausfüllen von Lücken in einem Text mit maskierten Wörtern.
+* Zusammenfassung: Erstellung einer Zusammenfassung einer langen Text- oder Dokumentensequenz.
+* Übersetzung: Übersetzen eines Textes in eine andere Sprache.
+* Merkmalsextraktion: Erstellen einer Tensordarstellung des Textes.
+
+**Bild**:
+* Bildklassifizierung: Klassifizierung eines Bildes.
+* Bildsegmentierung: Klassifizierung jedes Pixels in einem Bild.
+* Objekterkennung: Erkennen von Objekten innerhalb eines Bildes.
+
+**Audio**:
+* Audioklassifizierung: Zuweisung eines Labels zu einem bestimmten Audiosegment.
+* Automatische Spracherkennung (ASR): Transkription von Audiodaten in Text.
+
+<Tip>
+
+Für mehr Details über die [`pipeline`] und assoziierte Aufgaben, schauen Sie in die Dokumentation [hier](./main_classes/pipelines).
+
+</Tip>
+
+### Verwendung der Pipeline
+
+Im folgenden Beispiel werden Sie die [`pipeline`] für die Stimmungsanalyse verwenden.
+
+Installieren Sie die folgenden Abhängigkeiten, falls Sie dies nicht bereits getan haben:
+
+<frameworkcontent>
+<pt>
+```bash
+pip install torch
+```
+</pt>
+<tf>
+```bash
+pip install tensorflow
+```
+</tf>
+</frameworkcontent>
+
+Importieren sie die [`pipeline`] und spezifizieren sie die Aufgabe, welche sie lösen möchten:
+
+```py
+>>> from transformers import pipeline
+
+>>> classifier = pipeline("sentiment-analysis")
+```
+
+Die Pipeline lädt ein standardmäßiges [vortrainiertes Modell] (https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english) und einen Tokenizer für die Stimmungs-Analyse herunter und speichert sie. Jetzt können Sie den "Klassifikator" auf Ihren Zieltext anwenden:
+
+```py
+>>> classifier("We are very happy to show you the 🤗 Transformers library.")
+[{'label': 'POSITIVE', 'score': 0.9998}]
+```
+
+For more than one sentence, pass a list of sentences to the [`pipeline`] which returns a list of dictionaries:
+
+```py
+>>> results = classifier(["We are very happy to show you the 🤗 Transformers library.", "We hope you don't hate it."])
+>>> for result in results:
+...     print(f"label: {result['label']}, with score: {round(result['score'], 4)}")
+label: POSITIVE, with score: 0.9998
+label: NEGATIVE, with score: 0.5309
+```
+
+Die [`pipeline`] kann auch über einen ganzen Datensatz iterieren. Starten wir mit der Installation der [🤗 Datasets](https://huggingface.co/docs/datasets/) Bibliothek:
+
+```bash
+pip install datasets 
+```
+
+Erstellen wir eine [`pipeline`] mit der Aufgabe die wir lösen und dem Modell welches wir nutzen möchten.
+
+```py
+>>> import torch
+>>> from transformers import pipeline
+
+>>> speech_recognizer = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")
+```
+
+Als nächstes laden wir den Datensatz (siehe 🤗 Datasets [Quick Start](https://huggingface.co/docs/datasets/quickstart.html) für mehr Details) welches wir nutzen möchten. Zum Beispiel laden wir den [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) Datensatz:
+
+```py
+>>> from datasets import load_dataset, Audio
+
+>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")  # doctest: +IGNORE_RESULT
+```
+
+Wir müssen sicherstellen, dass die Abtastrate des Datensatzes der Abtastrate entspricht, mit der `facebook/wav2vec2-base-960h` trainiert wurde.
+
+```py
+>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=speech_recognizer.feature_extractor.sampling_rate))
+```
+
+Audiodateien werden automatisch geladen und neu abgetastet, wenn die Spalte "audio" aufgerufen wird.
+Extrahieren wir die rohen Wellenform-Arrays der ersten 4 Beispiele und übergeben wir sie als Liste an die Pipeline:
+
+```py
+>>> result = speech_recognizer(dataset[:4]["audio"])
+>>> print([d["text"] for d in result])
+['I WOULD LIKE TO SET UP A JOINT ACCOUNT WITH MY PARTNER HOW DO I PROCEED WITH DOING THAT', "FODING HOW I'D SET UP A JOIN TO HET WITH MY WIFE AND WHERE THE AP MIGHT BE", "I I'D LIKE TOY SET UP A JOINT ACCOUNT WITH MY PARTNER I'M NOT SEEING THE OPTION TO DO IT ON THE AP SO I CALLED IN TO GET SOME HELP CAN I JUST DO IT OVER THE PHONE WITH YOU AND GIVE YOU THE INFORMATION OR SHOULD I DO IT IN THE AP AND I'M MISSING SOMETHING UQUETTE HAD PREFERRED TO JUST DO IT OVER THE PHONE OF POSSIBLE THINGS", 'HOW DO I THURN A JOIN A COUNT']
+```
+
+Bei einem größeren Datensatz mit vielen Eingaben (wie bei Sprache oder Bildverarbeitung) sollten Sie einen Generator anstelle einer Liste übergeben, der alle Eingaben in den Speicher lädt. Weitere Informationen finden Sie in der [Pipeline-Dokumentation](./main_classes/pipelines).
+
+### Ein anderes Modell und einen anderen Tokenizer in der Pipeline verwenden
+
+Die [`pipeline`] kann jedes Modell aus dem [Model Hub] (https://huggingface.co/models) verwenden, wodurch es einfach ist, die [`pipeline`] für andere Anwendungsfälle anzupassen. Wenn Sie beispielsweise ein Modell wünschen, das französischen Text verarbeiten kann, verwenden Sie die Tags im Model Hub, um nach einem geeigneten Modell zu filtern. Das oberste gefilterte Ergebnis liefert ein mehrsprachiges [BERT-Modell](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment), das auf die Stimmungsanalyse abgestimmt ist. Großartig, verwenden wir dieses Modell!
+
+```py
+>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
+```
+
+<frameworkcontent>
+<pt>
+Use the [`AutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `AutoClass` below):
+
+```py
+>>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
+
+>>> model = AutoModelForSequenceClassification.from_pretrained(model_name)
+>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
+```
+</pt>
+<tf>
+Use the [`TFAutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `TFAutoClass` below):
+
+```py
+>>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
+
+>>> model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
+>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
+```
+</tf>
+</frameworkcontent>
+
+Dann können Sie das Modell und den Tokenizer in der [`pipeline`] angeben und den `Klassifikator` auf Ihren Zieltext anwenden:
+
+```py
+>>> classifier = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
+>>> classifier("Nous sommes très heureux de vous présenter la bibliothèque 🤗 Transformers.")
+[{'label': '5 stars', 'score': 0.7273}]
+```
+
+Wenn Sie kein Modell für Ihren Anwendungsfall finden können, müssen Sie ein vortrainiertes Modell auf Ihren Daten feinabstimmen. Schauen Sie sich unser [Feinabstimmungs-Tutorial](./training) an, um zu erfahren, wie das geht. Und schließlich, nachdem Sie Ihr trainiertes Modell verfeinert haben, sollten Sie es mit der Community im Model Hub teilen (siehe Tutorial [hier](./model_sharing)), um NLP für alle zu demokratisieren! 🤗
+
+## AutoClass
+
+<Youtube id="AhChOFRegn4"/>
+
+Unter der Haube arbeiten die Klassen [`AutoModelForSequenceClassification`] und [`AutoTokenizer`] zusammen, um die [`pipeline`] zu betreiben. Eine [`AutoClass`](./model_doc/auto) ist eine Abkürzung, die automatisch die Architektur eines trainierten Modells aus dessen Namen oder Pfad abruft. Sie müssen nur die passende `AutoClass` für Ihre Aufgabe und den zugehörigen Tokenizer mit [`AutoTokenizer`] auswählen. 
+
+Kehren wir zu unserem Beispiel zurück und sehen wir uns an, wie Sie die `AutoClass` verwenden können, um die Ergebnisse der [`pipeline`] zu replizieren.
+
+### AutoTokenizer
+
+Ein Tokenizer ist für die Vorverarbeitung von Text in ein für das Modell verständliches Format zuständig. Zunächst zerlegt der Tokenisierer den Text in Wörter, die *Token* genannt werden. Es gibt mehrere Regeln für den Tokenisierungsprozess, z. B. wie und auf welcher Ebene ein Wort aufgespalten wird (weitere Informationen über Tokenisierung [hier](./tokenizer_summary)). Das Wichtigste ist jedoch, dass Sie den Tokenizer mit demselben Modellnamen instanziieren müssen, um sicherzustellen, dass Sie dieselben Tokenisierungsregeln verwenden, mit denen ein Modell zuvor trainiert wurde.
+Laden sie einen Tokenizer mit [`AutoTokenizer`]:
+
+```py
+>>> from transformers import AutoTokenizer
+
+>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
+>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
+```
+
+Anschließend wandelt der Tokenizer die Token in Zahlen um, um einen Tensor als Eingabe für das Modell zu konstruieren. Dieser wird als *Vokabular* des Modells bezeichnet.
+
+Übergeben Sie Ihren Text an den Tokenizer:
+
+```py
+>>> encoding = tokenizer("We are very happy to show you the 🤗 Transformers library.")
+>>> print(encoding)
+{'input_ids': [101, 11312, 10320, 12495, 19308, 10114, 11391, 10855, 10103, 100, 58263, 13299, 119, 102],
+ 'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
+ 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
+```
+
+Der Tokenizer gibt ein Wörterbuch zurück, das Folgendes enthält:
+
+* [input_ids](./glossary#input-ids): numerische Repräsentationen Ihrer Token.
+* [atttention_mask](.glossary#attention-mask): gibt an, welche Token beachtet werden sollen.
+
+Genau wie die [`pipeline`] akzeptiert der Tokenizer eine Liste von Eingaben. Darüber hinaus kann der Tokenizer den Text auch auffüllen und kürzen, um einen Stapel mit einheitlicher Länge zurückzugeben:
+
+<frameworkcontent>
+<pt>
+```py
+>>> pt_batch = tokenizer(
+...     ["We are very happy to show you the 🤗 Transformers library.", "We hope you don't hate it."],
+...     padding=True,
+...     truncation=True,
+...     max_length=512,
+...     return_tensors="pt",
+... )
+```
+</pt>
+<tf>
+```py
+>>> tf_batch = tokenizer(
+...     ["We are very happy to show you the 🤗 Transformers library.", "We hope you don't hate it."],
+...     padding=True,
+...     truncation=True,
+...     max_length=512,
+...     return_tensors="tf",
+... )
+```
+</tf>
+</frameworkcontent>
+
+Lesen Sie das Tutorial [preprocessing](./preprocessing) für weitere Details zur Tokenisierung.
+
+### AutoModel
+
+<frameworkcontent>
+<pt>
+🤗 Transformers bietet eine einfache und einheitliche Möglichkeit, vortrainierte Instanzen zu laden. Das bedeutet, dass Sie ein [`AutoModel`] laden können, wie Sie einen [`AutoTokenizer`] laden würden. Der einzige Unterschied ist die Auswahl des richtigen [`AutoModel`] für die Aufgabe. Da Sie eine Text- oder Sequenzklassifizierung vornehmen, laden Sie [`AutoModelForSequenceClassification`]:
+
+```py
+>>> from transformers import AutoModelForSequenceClassification
+
+>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
+>>> pt_model = AutoModelForSequenceClassification.from_pretrained(model_name)
+```
+
+<Tip>
+
+In der [Aufgabenzusammenfassung](./task_summary) steht, welche [AutoModel]-Klasse für welche Aufgabe zu verwenden ist.
+
+</Tip>
+
+Jetzt können Sie Ihren vorverarbeiteten Stapel von Eingaben direkt an das Modell übergeben. Sie müssen nur das Wörterbuch entpacken, indem Sie `**` hinzufügen:
+
+```py
+>>> pt_outputs = pt_model(**pt_batch)
+```
+
+Das Modell gibt die endgültigen Aktivierungen in dem Attribut "logits" aus. Wenden Sie die Softmax-Funktion auf die "logits" an, um die Wahrscheinlichkeiten zu erhalten:
+  
+```py
+>>> from torch import nn
+
+>>> pt_predictions = nn.functional.softmax(pt_outputs.logits, dim=-1)
+>>> print(pt_predictions)
+tensor([[0.0021, 0.0018, 0.0115, 0.2121, 0.7725],
+        [0.2084, 0.1826, 0.1969, 0.1755, 0.2365]], grad_fn=<SoftmaxBackward0>)
+```
+</pt>
+<tf>
+🤗 Transformers bietet eine einfache und einheitliche Methode zum Laden von vortrainierten Instanzen. Das bedeutet, dass Sie ein [`TFAutoModel`] genauso laden können, wie Sie einen [`AutoTokenizer`] laden würden. Der einzige Unterschied ist die Auswahl des richtigen [`TFAutoModel`] für die Aufgabe. Da Sie Text - oder Sequenz - Klassifizierung machen, laden Sie [`TFAutoModelForSequenceClassification`]:
+
+```py
+>>> from transformers import TFAutoModelForSequenceClassification
+
+>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
+>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
+```
+
+<Tip>
+
+In der [Aufgabenzusammenfassung](./task_summary) steht, welche [AutoModel]-Klasse für welche Aufgabe zu verwenden ist.
+
+</Tip>
+
+Jetzt können Sie Ihren vorverarbeiteten Stapel von Eingaben direkt an das Modell übergeben, indem Sie die Wörterbuchschlüssel direkt an die Tensoren übergeben:
+  
+```py
+>>> tf_outputs = tf_model(tf_batch)
+```
+
+Das Modell gibt die endgültigen Aktivierungen in dem Attribut "logits" aus. Wenden Sie die Softmax-Funktion auf die "logits" an, um die Wahrscheinlichkeiten zu erhalten:
+
+```py
+>>> import tensorflow as tf
+
+>>> tf_predictions = tf.nn.softmax(tf_outputs.logits, axis=-1)
+>>> tf_predictions  # doctest: +IGNORE_RESULT
+```
+</tf>
+</frameworkcontent>
+
+<Tip>
+
+Alle 🤗 Transformers-Modelle (PyTorch oder TensorFlow) geben die Tensoren *vor* der endgültigen Aktivierungsfunktion
+Funktion (wie Softmax) aus, da die endgültige Aktivierungsfunktion oft mit dem Verlusten verschmolzen ist.
+
+</Tip>
+
+Modelle sind ein standardmäßiges [`torch.nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) oder ein [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model), sodass Sie sie in Ihrer üblichen Trainingsschleife verwenden können. Um jedoch die Dinge einfacher zu machen, bietet 🤗 Transformers eine [`Trainer`]-Klasse für PyTorch, die Funktionalität für verteiltes Training, gemischte Präzision und mehr bietet. Für TensorFlow können Sie die Methode `fit` aus [Keras](https://keras.io/) verwenden. Siehe das [training tutorial](./training) für weitere Details.
+
+<Tip>
+
+Transformers-Modellausgaben sind spezielle Datenklassen, so dass ihre Attribute in einer IDE automatisch vervollständigt werden.
+Die Modellausgänge verhalten sich auch wie ein Tupel oder ein Wörterbuch (z.B. können Sie mit einem Integer, einem Slice oder einem String indexieren), wobei die Attribute, die "None" sind, ignoriert werden.
+
+</Tip>
+
+### Modell speichern
+
+<frameworkcontent>
+<pt>
+Sobald Ihr Modell feinabgestimmt ist, können Sie es mit seinem Tokenizer speichern, indem Sie [`PreTrainedModel.save_pretrained`] verwenden:
+
+```py
+>>> pt_save_directory = "./pt_save_pretrained"
+>>> tokenizer.save_pretrained(pt_save_directory)  # doctest: +IGNORE_RESULT
+>>> pt_model.save_pretrained(pt_save_directory)
+```
+
+Wenn Sie bereit sind, das Modell erneut zu verwenden, laden Sie es mit [`PreTrainedModel.from_pretrained`]:
+
+```py
+>>> pt_model = AutoModelForSequenceClassification.from_pretrained("./pt_save_pretrained")
+```
+</pt>
+<tf>
+Sobald Ihr Modell feinabgestimmt ist, können Sie es mit seinem Tokenizer unter Verwendung von [`TFPreTrainedModel.save_pretrained`] speichern:
+
+```py
+>>> tf_save_directory = "./tf_save_pretrained"
+>>> tokenizer.save_pretrained(tf_save_directory)  # doctest: +IGNORE_RESULT
+>>> tf_model.save_pretrained(tf_save_directory)
+```
+
+Wenn Sie bereit sind, das Modell wieder zu verwenden, laden Sie es mit [`TFPreTrainedModel.from_pretrained`]:
+
+```py
+>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained("./tf_save_pretrained")
+```
+</tf>
+</frameworkcontent>
+
+Ein besonders cooles 🤗 Transformers-Feature ist die Möglichkeit, ein Modell zu speichern und es entweder als PyTorch- oder TensorFlow-Modell wieder zu laden. Der Parameter "from_pt" oder "from_tf" kann das Modell von einem Framework in das andere konvertieren:
+
+<frameworkcontent>
+<pt>
+```py
+>>> from transformers import AutoModel
+
+>>> tokenizer = AutoTokenizer.from_pretrained(tf_save_directory)
+>>> pt_model = AutoModelForSequenceClassification.from_pretrained(tf_save_directory, from_tf=True)
+```
+</pt>
+<tf>
+```py
+>>> from transformers import TFAutoModel
+
+>>> tokenizer = AutoTokenizer.from_pretrained(pt_save_directory)
+>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(pt_save_directory, from_pt=True)
+```
+</tf>
+</frameworkcontent>
+
+## Custom model builds
+
+Sie können die Konfigurationsklasse des Modells ändern, um zu bestimmen, wie ein Modell aufgebaut ist. Die Konfiguration legt die Attribute eines Modells fest, z. B. die Anzahl der verborgenen Schichten oder der Aufmerksamkeitsköpfe. Wenn Sie ein Modell aus einer benutzerdefinierten Konfigurationsklasse initialisieren, beginnen Sie bei Null. Die Modellattribute werden zufällig initialisiert, und Sie müssen das Modell trainieren, bevor Sie es verwenden können, um aussagekräftige Ergebnisse zu erhalten.
+
+Beginnen Sie mit dem Import von [`AutoConfig`] und laden Sie dann das trainierte Modell, das Sie ändern möchten. Innerhalb von [`AutoConfig.from_pretrained`] können Sie das Attribut angeben, das Sie ändern möchten, z. B. die Anzahl der Aufmerksamkeitsköpfe:
+
+```py
+>>> from transformers import AutoConfig
+
+>>> my_config = AutoConfig.from_pretrained("distilbert-base-uncased", n_heads=12)
+```
+
+<frameworkcontent>
+<pt>
+Create a model from your custom configuration with [`AutoModel.from_config`]:
+
+```py
+>>> from transformers import AutoModel
+
+>>> my_model = AutoModel.from_config(my_config)
+```
+</pt>
+<tf>
+Create a model from your custom configuration with [`TFAutoModel.from_config`]:
+
+```py
+>>> from transformers import TFAutoModel
+
+>>> my_model = TFAutoModel.from_config(my_config)
+```
+</tf>
+</frameworkcontent>
+
+Weitere Informationen zur Erstellung von benutzerdefinierten Konfigurationen finden Sie in der Anleitung [Erstellen einer benutzerdefinierten Architektur](./create_a_model).
+
+## Wie geht es weiter?
+
+Nachdem Sie nun die 🤗 Transformers-Kurztour abgeschlossen haben, schauen Sie sich unsere Anleitungen an und erfahren Sie, wie Sie spezifischere Dinge tun können, wie das Schreiben eines benutzerdefinierten Modells, die Feinabstimmung eines Modells für eine Aufgabe und wie man ein Modell mit einem Skript trainiert. Wenn Sie mehr über die Kernkonzepte von 🤗 Transformers erfahren möchten, nehmen Sie sich eine Tasse Kaffee und werfen Sie einen Blick auf unsere konzeptionellen Leitfäden!

docs/source/en/_toctree.yml

@@ -1,4 +1,4 @@
-- sections: 
+- sections:
   - local: index
     title: 🤗 Transformers
   - local: quicktour
@@ -21,72 +21,107 @@
     title: Share a model
   title: Tutorials
 - sections:
-  - local: fast_tokenizers
-    title: "Use tokenizers from 🤗 Tokenizers"
-  - local: create_a_model
-    title: Create a custom architecture
-  - local: custom_models
-    title: Sharing custom models
   - sections:
-    - local: tasks/sequence_classification
-      title: Text classification
-    - local: tasks/token_classification
-      title: Token classification
-    - local: tasks/question_answering
-      title: Question answering
-    - local: tasks/language_modeling
-      title: Language modeling
-    - local: tasks/translation
-      title: Translation
-    - local: tasks/summarization
-      title: Summarization
-    - local: tasks/multiple_choice
-      title: Multiple choice
+    - local: create_a_model
+      title: Create a custom architecture
+    - local: custom_models
+      title: Sharing custom models
+    - local: run_scripts
+      title: Train with a script
+    - local: sagemaker
+      title: Run training on Amazon SageMaker
+    - local: converting_tensorflow_models
+      title: Converting TensorFlow Checkpoints
+    - local: serialization
+      title: Export 🤗 Transformers models
+    - local: troubleshooting
+      title: Troubleshoot
+    title: General usage
+  - sections:
+    - local: fast_tokenizers
+      title: Use tokenizers from 🤗 Tokenizers
+    - local: multilingual
+      title: Inference for multilingual models
+    - sections:
+      - local: tasks/sequence_classification
+        title: Text classification
+      - local: tasks/token_classification
+        title: Token classification
+      - local: tasks/question_answering
+        title: Question answering
+      - local: tasks/language_modeling
+        title: Language modeling
+      - local: tasks/translation
+        title: Translation
+      - local: tasks/summarization
+        title: Summarization
+      - local: tasks/multiple_choice
+        title: Multiple choice
+      title: Task guides
+      isExpanded: false
+    title: Natural Language Processing
+  - sections:
     - local: tasks/audio_classification
       title: Audio classification
     - local: tasks/asr
       title: Automatic speech recognition
+    title: Audio
+  - sections:
     - local: tasks/image_classification
       title: Image classification
-    title: Fine-tune for downstream tasks
-  - local: run_scripts
-    title: Train with a script
-  - local: sagemaker
-    title: Run training on Amazon SageMaker
-  - local: multilingual
-    title: Inference for multilingual models
-  - local: converting_tensorflow_models
-    title: Converting TensorFlow Checkpoints
-  - local: serialization
-    title: Export 🤗 Transformers models
-  - local: performance
-    title: 'Performance and Scalability: How To Fit a Bigger Model and Train It Faster'
-  - local: big_models
-    title: Instantiating a big model
-  - local: parallelism
-    title: Model Parallelism
+    - local: tasks/semantic_segmentation
+      title: Semantic segmentation
+    title: Computer Vision
+  - sections:
+    - local: performance
+      title: Overview
+    - local: perf_train_gpu_one
+      title: Training on one GPU
+    - local: perf_train_gpu_many
+      title: Training on many GPUs
+    - local: perf_train_cpu
+      title: Training on CPU
+    - local: perf_train_cpu_many
+      title: Training on many CPUs
+    - local: perf_train_tpu
+      title: Training on TPUs
+    - local: perf_train_special
+      title: Training on Specialized Hardware
+    - local: perf_infer_cpu
+      title: Inference on CPU
+    - local: perf_infer_gpu_one
+      title: Inference on one GPU
+    - local: perf_infer_gpu_many
+      title: Inference on many GPUs
+    - local: perf_infer_special
+      title: Inference on Specialized Hardware
+    - local: perf_hardware
+      title: Custom hardware for training
+    - local: big_models
+      title: Instantiating a big model
+    - local: debugging
+      title: Debugging
+    title: Performance and scalability
+  - sections:
+    - local: contributing
+      title: How to contribute to transformers?
+    - local: add_new_model
+      title: How to add a model to 🤗 Transformers?
+    - local: add_new_pipeline
+      title: How to add a pipeline to 🤗 Transformers?
+    - local: testing
+      title: Testing
+    - local: pr_checks
+      title: Checks on a Pull Request
+    title: Contribute
+  - local: notebooks
+    title: 🤗 Transformers Notebooks
+  - local: community
+    title: Community resources
   - local: benchmarks
     title: Benchmarks
   - local: migration
     title: Migrating from previous packages
-  - local: troubleshooting
-    title: Troubleshoot
-  - local: debugging
-    title: Debugging
-  - local: notebooks
-    title: "🤗 Transformers Notebooks"
-  - local: community
-    title: Community
-  - local: contributing
-    title: How to contribute to transformers?
-  - local: add_new_model
-    title: "How to add a model to 🤗 Transformers?"
-  - local: add_new_pipeline
-    title: "How to add a pipeline to 🤗 Transformers?"
-  - local: testing
-    title: Testing
-  - local: pr_checks
-    title: Checks on a Pull Request
   title: How-to guides
 - sections:
   - local: philosophy
@@ -142,250 +177,311 @@
       title: Feature Extractor
     title: Main Classes
   - sections:
-    - local: model_doc/albert
-      title: ALBERT
     - local: model_doc/auto
       title: Auto Classes
-    - local: model_doc/bart
-      title: BART
-    - local: model_doc/barthez
-      title: BARThez
-    - local: model_doc/bartpho
-      title: BARTpho
-    - local: model_doc/beit
-      title: BEiT
-    - local: model_doc/bert
-      title: BERT
-    - local: model_doc/bertweet
-      title: Bertweet
-    - local: model_doc/bert-generation
-      title: BertGeneration
-    - local: model_doc/bert-japanese
-      title: BertJapanese
-    - local: model_doc/big_bird
-      title: BigBird
-    - local: model_doc/bigbird_pegasus
-      title: BigBirdPegasus
-    - local: model_doc/blenderbot
-      title: Blenderbot
-    - local: model_doc/blenderbot-small
-      title: Blenderbot Small
-    - local: model_doc/bort
-      title: BORT
-    - local: model_doc/byt5
-      title: ByT5
-    - local: model_doc/camembert
-      title: CamemBERT
-    - local: model_doc/canine
-      title: CANINE
-    - local: model_doc/convnext
-      title: ConvNeXT
-    - local: model_doc/clip
-      title: CLIP
-    - local: model_doc/convbert
-      title: ConvBERT
-    - local: model_doc/cpm
-      title: CPM
-    - local: model_doc/ctrl
-      title: CTRL
-    - local: model_doc/data2vec
-      title: Data2Vec
-    - local: model_doc/deberta
-      title: DeBERTa
-    - local: model_doc/deberta-v2
-      title: DeBERTa-v2
-    - local: model_doc/decision_transformer
-      title: Decision Transformer
-    - local: model_doc/deit
-      title: DeiT
-    - local: model_doc/detr
-      title: DETR
-    - local: model_doc/dialogpt
-      title: DialoGPT
-    - local: model_doc/distilbert
-      title: DistilBERT
-    - local: model_doc/dit
-      title: DiT
-    - local: model_doc/dpr
-      title: DPR
-    - local: model_doc/dpt
-      title: DPT
-    - local: model_doc/electra
-      title: ELECTRA
-    - local: model_doc/encoder-decoder
-      title: Encoder Decoder Models
-    - local: model_doc/flaubert
-      title: FlauBERT
-    - local: model_doc/flava
-      title: FLAVA
-    - local: model_doc/fnet
-      title: FNet
-    - local: model_doc/fsmt
-      title: FSMT
-    - local: model_doc/funnel
-      title: Funnel Transformer
-    - local: model_doc/glpn
-      title: GLPN
-    - local: model_doc/herbert
-      title: HerBERT
-    - local: model_doc/ibert
-      title: I-BERT
-    - local: model_doc/imagegpt
-      title: ImageGPT
-    - local: model_doc/layoutlm
-      title: LayoutLM
-    - local: model_doc/layoutlmv2
-      title: LayoutLMV2
-    - local: model_doc/layoutxlm
-      title: LayoutXLM
-    - local: model_doc/led
-      title: LED
-    - local: model_doc/longformer
-      title: Longformer
-    - local: model_doc/luke
-      title: LUKE
-    - local: model_doc/lxmert
-      title: LXMERT
-    - local: model_doc/marian
-      title: MarianMT
-    - local: model_doc/maskformer
-      title: MaskFormer
-    - local: model_doc/m2m_100
-      title: M2M100
-    - local: model_doc/mbart
-      title: MBart and MBart-50
-    - local: model_doc/megatron-bert
-      title: MegatronBERT
-    - local: model_doc/megatron_gpt2
-      title: MegatronGPT2
-    - local: model_doc/mluke
-      title: mLUKE
-    - local: model_doc/mobilebert
-      title: MobileBERT
-    - local: model_doc/mpnet
-      title: MPNet
-    - local: model_doc/mt5
-      title: MT5
-    - local: model_doc/nystromformer
-      title: Nyströmformer
-    - local: model_doc/openai-gpt
-      title: OpenAI GPT
-    - local: model_doc/opt
-      title: OPT
-    - local: model_doc/gpt2
-      title: OpenAI GPT2
-    - local: model_doc/gptj
-      title: GPT-J
-    - local: model_doc/gpt_neo
-      title: GPT Neo
-    - local: model_doc/hubert
-      title: Hubert
-    - local: model_doc/perceiver
-      title: Perceiver
-    - local: model_doc/pegasus
-      title: Pegasus
-    - local: model_doc/phobert
-      title: PhoBERT
-    - local: model_doc/plbart
-      title: PLBart
-    - local: model_doc/poolformer
-      title: PoolFormer
-    - local: model_doc/prophetnet
-      title: ProphetNet
-    - local: model_doc/qdqbert
-      title: QDQBert
-    - local: model_doc/rag
-      title: RAG
-    - local: model_doc/realm
-      title: REALM
-    - local: model_doc/reformer
-      title: Reformer
-    - local: model_doc/rembert
-      title: RemBERT
-    - local: model_doc/regnet
-      title: RegNet
-    - local: model_doc/resnet
-      title: ResNet
-    - local: model_doc/retribert
-      title: RetriBERT
-    - local: model_doc/roberta
-      title: RoBERTa
-    - local: model_doc/roformer
-      title: RoFormer
-    - local: model_doc/segformer
-      title: SegFormer
-    - local: model_doc/sew
-      title: SEW
-    - local: model_doc/sew-d
-      title: SEW-D
-    - local: model_doc/speech-encoder-decoder
-      title: Speech Encoder Decoder Models
-    - local: model_doc/speech_to_text
-      title: Speech2Text
-    - local: model_doc/speech_to_text_2
-      title: Speech2Text2
-    - local: model_doc/splinter
-      title: Splinter
-    - local: model_doc/squeezebert
-      title: SqueezeBERT
-    - local: model_doc/swin
-      title: Swin Transformer
-    - local: model_doc/t5
-      title: T5
-    - local: model_doc/t5v1.1
-      title: T5v1.1
-    - local: model_doc/tapas
-      title: TAPAS
-    - local: model_doc/tapex
-      title: TAPEX
-    - local: model_doc/transfo-xl
-      title: Transformer XL
-    - local: model_doc/trocr
-      title: TrOCR
-    - local: model_doc/unispeech
-      title: UniSpeech
-    - local: model_doc/unispeech-sat
-      title: UniSpeech-SAT
-    - local: model_doc/van
-      title: VAN
-    - local: model_doc/vilt
-      title: ViLT
-    - local: model_doc/vision-encoder-decoder
-      title: Vision Encoder Decoder Models
-    - local: model_doc/vision-text-dual-encoder
-      title: Vision Text Dual Encoder
-    - local: model_doc/vit
-      title: Vision Transformer (ViT)
-    - local: model_doc/vit_mae
-      title: ViTMAE
-    - local: model_doc/visual_bert
-      title: VisualBERT
-    - local: model_doc/wav2vec2
-      title: Wav2Vec2
-    - local: model_doc/wav2vec2_phoneme
-      title: Wav2Vec2Phoneme
-    - local: model_doc/wavlm
-      title: WavLM
-    - local: model_doc/xglm
-      title: XGLM
-    - local: model_doc/xlm
-      title: XLM
-    - local: model_doc/xlm-prophetnet
-      title: XLM-ProphetNet
-    - local: model_doc/xlm-roberta
-      title: XLM-RoBERTa
-    - local: model_doc/xlm-roberta-xl
-      title: XLM-RoBERTa-XL
-    - local: model_doc/xlnet
-      title: XLNet
-    - local: model_doc/xlsr_wav2vec2
-      title: XLSR-Wav2Vec2
-    - local: model_doc/xls_r
-      title: XLS-R
-    - local: model_doc/yolos
-      title: YOLOS
-    - local: model_doc/yoso
-      title: YOSO
+    - isExpanded: false
+      sections:
+      - local: model_doc/albert
+        title: ALBERT
+      - local: model_doc/bart
+        title: BART
+      - local: model_doc/barthez
+        title: BARThez
+      - local: model_doc/bartpho
+        title: BARTpho
+      - local: model_doc/bert
+        title: BERT
+      - local: model_doc/bert-generation
+        title: BertGeneration
+      - local: model_doc/bert-japanese
+        title: BertJapanese
+      - local: model_doc/bertweet
+        title: Bertweet
+      - local: model_doc/big_bird
+        title: BigBird
+      - local: model_doc/bigbird_pegasus
+        title: BigBirdPegasus
+      - local: model_doc/blenderbot
+        title: Blenderbot
+      - local: model_doc/blenderbot-small
+        title: Blenderbot Small
+      - local: model_doc/bloom
+        title: BLOOM
+      - local: model_doc/bort
+        title: BORT
+      - local: model_doc/byt5
+        title: ByT5
+      - local: model_doc/camembert
+        title: CamemBERT
+      - local: model_doc/canine
+        title: CANINE
+      - local: model_doc/codegen
+        title: CodeGen
+      - local: model_doc/convbert
+        title: ConvBERT
+      - local: model_doc/cpm
+        title: CPM
+      - local: model_doc/ctrl
+        title: CTRL
+      - local: model_doc/deberta
+        title: DeBERTa
+      - local: model_doc/deberta-v2
+        title: DeBERTa-v2
+      - local: model_doc/dialogpt
+        title: DialoGPT
+      - local: model_doc/distilbert
+        title: DistilBERT
+      - local: model_doc/dpr
+        title: DPR
+      - local: model_doc/electra
+        title: ELECTRA
+      - local: model_doc/encoder-decoder
+        title: Encoder Decoder Models
+      - local: model_doc/ernie
+        title: ERNIE
+      - local: model_doc/flaubert
+        title: FlauBERT
+      - local: model_doc/fnet
+        title: FNet
+      - local: model_doc/fsmt
+        title: FSMT
+      - local: model_doc/funnel
+        title: Funnel Transformer
+      - local: model_doc/openai-gpt
+        title: GPT
+      - local: model_doc/gpt_neo
+        title: GPT Neo
+      - local: model_doc/gpt_neox
+        title: GPT NeoX
+      - local: model_doc/gptj
+        title: GPT-J
+      - local: model_doc/gpt2
+        title: GPT2
+      - local: model_doc/herbert
+        title: HerBERT
+      - local: model_doc/ibert
+        title: I-BERT
+      - local: model_doc/layoutlm
+        title: LayoutLM
+      - local: model_doc/led
+        title: LED
+      - local: model_doc/longformer
+        title: Longformer
+      - local: model_doc/longt5
+        title: LongT5
+      - local: model_doc/luke
+        title: LUKE
+      - local: model_doc/m2m_100
+        title: M2M100
+      - local: model_doc/marian
+        title: MarianMT
+      - local: model_doc/mbart
+        title: MBart and MBart-50
+      - local: model_doc/megatron-bert
+        title: MegatronBERT
+      - local: model_doc/megatron_gpt2
+        title: MegatronGPT2
+      - local: model_doc/mluke
+        title: mLUKE
+      - local: model_doc/mobilebert
+        title: MobileBERT
+      - local: model_doc/mpnet
+        title: MPNet
+      - local: model_doc/mt5
+        title: MT5
+      - local: model_doc/mvp
+        title: MVP
+      - local: model_doc/nezha
+        title: NEZHA
+      - local: model_doc/nllb
+        title: NLLB
+      - local: model_doc/nystromformer
+        title: Nyströmformer
+      - local: model_doc/opt
+        title: OPT
+      - local: model_doc/pegasus
+        title: Pegasus
+      - local: model_doc/pegasus_x
+        title: PEGASUS-X
+      - local: model_doc/phobert
+        title: PhoBERT
+      - local: model_doc/plbart
+        title: PLBart
+      - local: model_doc/prophetnet
+        title: ProphetNet
+      - local: model_doc/qdqbert
+        title: QDQBert
+      - local: model_doc/rag
+        title: RAG
+      - local: model_doc/realm
+        title: REALM
+      - local: model_doc/reformer
+        title: Reformer
+      - local: model_doc/rembert
+        title: RemBERT
+      - local: model_doc/retribert
+        title: RetriBERT
+      - local: model_doc/roberta
+        title: RoBERTa
+      - local: model_doc/roformer
+        title: RoFormer
+      - local: model_doc/splinter
+        title: Splinter
+      - local: model_doc/squeezebert
+        title: SqueezeBERT
+      - local: model_doc/t5
+        title: T5
+      - local: model_doc/t5v1.1
+        title: T5v1.1
+      - local: model_doc/tapas
+        title: TAPAS
+      - local: model_doc/tapex
+        title: TAPEX
+      - local: model_doc/transfo-xl
+        title: Transformer XL
+      - local: model_doc/ul2
+        title: UL2
+      - local: model_doc/xglm
+        title: XGLM
+      - local: model_doc/xlm
+        title: XLM
+      - local: model_doc/xlm-prophetnet
+        title: XLM-ProphetNet
+      - local: model_doc/xlm-roberta
+        title: XLM-RoBERTa
+      - local: model_doc/xlm-roberta-xl
+        title: XLM-RoBERTa-XL
+      - local: model_doc/xlnet
+        title: XLNet
+      - local: model_doc/yoso
+        title: YOSO
+      title: Text models
+    - isExpanded: false
+      sections:
+      - local: model_doc/beit
+        title: BEiT
+      - local: model_doc/convnext
+        title: ConvNeXT
+      - local: model_doc/cvt
+        title: CvT
+      - local: model_doc/deit
+        title: DeiT
+      - local: model_doc/detr
+        title: DETR
+      - local: model_doc/dit
+        title: DiT
+      - local: model_doc/dpt
+        title: DPT
+      - local: model_doc/glpn
+        title: GLPN
+      - local: model_doc/imagegpt
+        title: ImageGPT
+      - local: model_doc/levit
+        title: LeViT
+      - local: model_doc/maskformer
+        title: MaskFormer
+      - local: model_doc/mobilevit
+        title: MobileViT
+      - local: model_doc/poolformer
+        title: PoolFormer
+      - local: model_doc/regnet
+        title: RegNet
+      - local: model_doc/resnet
+        title: ResNet
+      - local: model_doc/segformer
+        title: SegFormer
+      - local: model_doc/swin
+        title: Swin Transformer
+      - local: model_doc/swinv2
+        title: Swin Transformer V2
+      - local: model_doc/van
+        title: VAN
+      - local: model_doc/videomae
+        title: VideoMAE
+      - local: model_doc/vit
+        title: Vision Transformer (ViT)
+      - local: model_doc/vit_mae
+        title: ViTMAE
+      - local: model_doc/yolos
+        title: YOLOS
+      title: Vision models
+    - isExpanded: false
+      sections:
+      - local: model_doc/hubert
+        title: Hubert
+      - local: model_doc/mctct
+        title: MCTCT
+      - local: model_doc/sew
+        title: SEW
+      - local: model_doc/sew-d
+        title: SEW-D
+      - local: model_doc/speech_to_text
+        title: Speech2Text
+      - local: model_doc/speech_to_text_2
+        title: Speech2Text2
+      - local: model_doc/unispeech
+        title: UniSpeech
+      - local: model_doc/unispeech-sat
+        title: UniSpeech-SAT
+      - local: model_doc/wav2vec2
+        title: Wav2Vec2
+      - local: model_doc/wav2vec2-conformer
+        title: Wav2Vec2-Conformer
+      - local: model_doc/wav2vec2_phoneme
+        title: Wav2Vec2Phoneme
+      - local: model_doc/wavlm
+        title: WavLM
+      - local: model_doc/xls_r
+        title: XLS-R
+      - local: model_doc/xlsr_wav2vec2
+        title: XLSR-Wav2Vec2
+      title: Audio models
+    - isExpanded: false
+      sections:
+      - local: model_doc/clip
+        title: CLIP
+      - local: model_doc/data2vec
+        title: Data2Vec
+      - local: model_doc/donut
+        title: Donut
+      - local: model_doc/flava
+        title: FLAVA
+      - local: model_doc/groupvit
+        title: GroupViT
+      - local: model_doc/layoutlmv2
+        title: LayoutLMV2
+      - local: model_doc/layoutlmv3
+        title: LayoutLMV3
+      - local: model_doc/layoutxlm
+        title: LayoutXLM
+      - local: model_doc/lxmert
+        title: LXMERT
+      - local: model_doc/owlvit
+        title: OWL-ViT
+      - local: model_doc/perceiver
+        title: Perceiver
+      - local: model_doc/speech-encoder-decoder
+        title: Speech Encoder Decoder Models
+      - local: model_doc/trocr
+        title: TrOCR
+      - local: model_doc/vilt
+        title: ViLT
+      - local: model_doc/vision-encoder-decoder
+        title: Vision Encoder Decoder Models
+      - local: model_doc/vision-text-dual-encoder
+        title: Vision Text Dual Encoder
+      - local: model_doc/visual_bert
+        title: VisualBERT
+      - local: model_doc/xclip
+        title: X-CLIP
+      title: Multimodal models
+    - isExpanded: false
+      sections:
+      - local: model_doc/decision_transformer
+        title: Decision Transformer
+      - local: model_doc/trajectory_transformer
+        title: Trajectory Transformer
+      title: Reinforcement learning models
     title: Models
   - sections:
     - local: internal/modeling_utils
@@ -401,4 +497,4 @@
     - local: internal/file_utils
       title: General Utilities
     title: Internal Helpers
-  title: API
+  title: API

docs/source/en/accelerate.mdx

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Distributed training with 🤗 Accelerate
 
-As models get bigger, parallelism has emerged as a strategy for training larger models on limited hardware and accelerating training speed by several orders of magnitude. At Hugging Face, we created the [🤗 Accelerate](https://huggingface.co/docs/accelerate/index.html) library to help users easily train a 🤗 Transformers model on any type of distributed setup, whether it is multiple GPU's on one machine or multiple GPU's across several machines. In this tutorial, learn how to customize your native PyTorch training loop to enable training in a distributed environment.
+As models get bigger, parallelism has emerged as a strategy for training larger models on limited hardware and accelerating training speed by several orders of magnitude. At Hugging Face, we created the [🤗 Accelerate](https://huggingface.co/docs/accelerate) library to help users easily train a 🤗 Transformers model on any type of distributed setup, whether it is multiple GPU's on one machine or multiple GPU's across several machines. In this tutorial, learn how to customize your native PyTorch training loop to enable training in a distributed environment.
 
 ## Setup
 
@@ -22,7 +22,7 @@ Get started by installing 🤗 Accelerate:
 pip install accelerate
 ```
 
-Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator) object. `Accelerator` will automatically detect your type of distributed setup and initialize all the necessary components for training. You don't need to explicitly place your model on a device.
+Then import and create an [`~accelerate.Accelerator`] object. The [`~accelerate.Accelerator`] will automatically detect your type of distributed setup and initialize all the necessary components for training. You don't need to explicitly place your model on a device.
 
 ```py
 >>> from accelerate import Accelerator
@@ -32,7 +32,7 @@ Then import and create an [`Accelerator`](https://huggingface.co/docs/accelerate
 
 ## Prepare to accelerate
 
-The next step is to pass all the relevant training objects to the [`prepare`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.prepare) method. This includes your training and evaluation DataLoaders, a model and an optimizer:
+The next step is to pass all the relevant training objects to the [`~accelerate.Accelerator.prepare`] method. This includes your training and evaluation DataLoaders, a model and an optimizer:
 
 ```py
 >>> train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
@@ -42,7 +42,7 @@ The next step is to pass all the relevant training objects to the [`prepare`](ht
 
 ## Backward
 
-The last addition is to replace the typical `loss.backward()` in your training loop with 🤗 Accelerate's [`backward`](https://huggingface.co/docs/accelerate/accelerator.html#accelerate.Accelerator.backward) method:
+The last addition is to replace the typical `loss.backward()` in your training loop with 🤗 Accelerate's [`~accelerate.Accelerator.backward`]method:
 
 ```py
 >>> for epoch in range(num_epochs):
@@ -121,7 +121,7 @@ accelerate launch train.py
 
 ### Train with a notebook
 
-🤗 Accelerate can also run in a notebook if you're planning on using Colaboratory's TPUs. Wrap all the code responsible for training in a function, and pass it to `notebook_launcher`:
+🤗 Accelerate can also run in a notebook if you're planning on using Colaboratory's TPUs. Wrap all the code responsible for training in a function, and pass it to [`~accelerate.notebook_launcher`]:
 
 ```py
 >>> from accelerate import notebook_launcher
@@ -129,4 +129,4 @@ accelerate launch train.py
 >>> notebook_launcher(training_function)
 ```
 
-For more information about 🤗 Accelerate and it's rich features, refer to the [documentation](https://huggingface.co/docs/accelerate/index.html).
+For more information about 🤗 Accelerate and it's rich features, refer to the [documentation](https://huggingface.co/docs/accelerate).

docs/source/en/add_new_model.mdx

@@ -813,13 +813,9 @@ checkpoint and to get the required access rights to be able to upload the model
 *brand_new_bert*. The `push_to_hub` method, present in all models in `transformers`, is a quick and efficient way to push your checkpoint to the hub. A little snippet is pasted below:
 
 ```python
-brand_new_bert.push_to_hub(
-    repo_path_or_name="brand_new_bert",
-    # Uncomment the following line to push to an organization
-    # organization="<ORGANIZATION>",
-    commit_message="Add model",
-    use_temp_dir=True,
-)
+brand_new_bert.push_to_hub("brand_new_bert")
+# Uncomment the following line to push to an organization.
+# brand_new_bert.push_to_hub("<organization>/brand_new_bert")
 ```
 
 It is worth spending some time to create fitting model cards for each checkpoint. The model cards should highlight the

docs/source/en/add_new_pipeline.mdx

@@ -9,7 +9,10 @@ Unless required by applicable law or agreed to in writing, software distributed
 an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
 -->
 
-# How to add a pipeline to 🤗 Transformers?
+# How to create a custom pipeline?
+
+In this guide, we will see how to create a custom pipeline and share it on the [Hub](hf.co/models) or add it to the
+Transformers library.
 
 First and foremost, you need to decide the raw entries the pipeline will be able to take. It can be strings, raw bytes,
 dictionaries or whatever seems to be the most likely desired input. Try to keep these inputs as pure Python as possible
@@ -99,7 +102,7 @@ def _sanitize_parameters(self, **kwargs):
 
     postprocess_kwargs = {}
     if "top_k" in kwargs:
-        preprocess_kwargs["top_k"] = kwargs["top_k"]
+        postprocess_kwargs["top_k"] = kwargs["top_k"]
     return preprocess_kwargs, {}, postprocess_kwargs
 ```
 
@@ -111,12 +114,123 @@ of arguments for ease of use (audio files, can be filenames, URLs or pure bytes)
 
 ## Adding it to the list of supported tasks
 
-Go to `src/transformers/pipelines/__init__.py` and fill in `SUPPORTED_TASKS` with your newly created pipeline.
-If possible it should provide a default model.
+To register your `new-task` to the list of supported tasks, you have to add it to the `PIPELINE_REGISTRY`:
 
-## Adding tests
+```python
+from transformers.pipelines import PIPELINE_REGISTRY
 
-Create a new file `tests/test_pipelines_MY_PIPELINE.py` with example with the other tests.
+PIPELINE_REGISTRY.register_pipeline(
+    "new-task",
+    pipeline_class=MyPipeline,
+    pt_model=AutoModelForSequenceClassification,
+)
+```
+
+You can specify a default model if you want, in which case it should come with a specific revision (which can be the name of a branch or a commit hash, here we took `"abcdef"`) as well was the type:
+
+```python
+PIPELINE_REGISTRY.register_pipeline(
+    "new-task",
+    pipeline_class=MyPipeline,
+    pt_model=AutoModelForSequenceClassification,
+    default={"pt": ("user/awesome_model", "abcdef")},
+    type="text",  # current support type: text, audio, image, multimodal
+)
+```
+
+## Share your pipeline on the Hub
+
+To share your custom pipeline on the Hub, you just have to save the custom code of your `Pipeline` subclass in a
+python file. For instance, let's say we want to use a custom pipeline for sentence pair classification like this:
+
+```py
+import numpy as np
+
+from transformers import Pipeline
+
+
+def softmax(outputs):
+    maxes = np.max(outputs, axis=-1, keepdims=True)
+    shifted_exp = np.exp(outputs - maxes)
+    return shifted_exp / shifted_exp.sum(axis=-1, keepdims=True)
+
+
+class PairClassificationPipeline(Pipeline):
+    def _sanitize_parameters(self, **kwargs):
+        preprocess_kwargs = {}
+        if "second_text" in kwargs:
+            preprocess_kwargs["second_text"] = kwargs["second_text"]
+        return preprocess_kwargs, {}, {}
+
+    def preprocess(self, text, second_text=None):
+        return self.tokenizer(text, text_pair=second_text, return_tensors=self.framework)
+
+    def _forward(self, model_inputs):
+        return self.model(**model_inputs)
+
+    def postprocess(self, model_outputs):
+        logits = model_outputs.logits[0].numpy()
+        probabilities = softmax(logits)
+
+        best_class = np.argmax(probabilities)
+        label = self.model.config.id2label[best_class]
+        score = probabilities[best_class].item()
+        logits = logits.tolist()
+        return {"label": label, "score": score, "logits": logits}
+```
+
+The implementation is framework agnostic, and will work for PyTorch and TensorFlow models. If we have saved this in
+a file named `pair_classification.py`, we can then import it and register it like this:
+
+```py
+from pair_classification import PairClassificationPipeline
+from transformers.pipelines import PIPELINE_REGISTRY
+from transformers import AutoModelForSequenceClassification, TFAutoModelForSequenceClassification
+
+PIPELINE_REGISTRY.register_pipeline(
+    "pair-classification",
+    pipeline_class=PairClassificationPipeline,
+    pt_model=AutoModelForSequenceClassification,
+    tf_model=TFAutoModelForSequenceClassification,
+)
+```
+
+Once this is done, we can use it with a pretrained model. For instance `sgugger/finetuned-bert-mrpc` has been
+fine-tuned on the MRPC dataset, which classifies pairs of sentences as paraphrases or not.
+
+```py
+from transformers import pipeline
+
+classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
+```
+
+Then we can share it on the Hub by using the `save_pretrained` method in a `Repository`:
+
+```py
+from huggingface_hub import Repository
+
+repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
+classifier.save_pretrained("test-dynamic-pipeline")
+repo.push_to_hub()
+```
+
+This will copy the file where you defined `PairClassificationPipeline` inside the folder `"test-dynamic-pipeline"`,
+along with saving the model and tokenizer of the pipeline, before pushing everything in the repository
+`{your_username}/test-dynamic-pipeline`. After that anyone can use it as long as they provide the option
+`trust_remote_code=True`:
+
+```py
+from transformers import pipeline
+
+classifier = pipeline(model="{your_username}/test-dynamic-pipeline", trust_remote_code=True)
+```
+
+## Add the pipeline to Transformers
+
+If you want to contribute your pipeline to Transformers, you will need to add a new module in the `pipelines` submodule
+with the code of your pipeline, then add it in the list of tasks defined in `pipelines/__init__.py`.
+
+Then you will need to add tests. Create a new file `tests/test_pipelines_MY_PIPELINE.py` with example with the other tests.
 
 The `run_pipeline_test` function will be very generic and run on small random models on every possible
 architecture as defined by `model_mapping` and `tf_model_mapping`.

docs/source/en/autoclass_tutorial.mdx

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Load pretrained instances with an AutoClass
 
-With so many different Transformer architectures, it can be challenging to create one for your checkpoint. As a part of 🤗 Transformers core philosophy to make the library easy, simple and flexible to use, an `AutoClass` automatically infer and load the correct architecture from a given checkpoint. The `from_pretrained` method lets you quickly load a pretrained model for any architecture so you don't have to devote time and resources to train a model from scratch. Producing this type of checkpoint-agnostic code means if your code works for one checkpoint, it will work with another checkpoint - as long as it was trained for a similar task - even if the architecture is different.
+With so many different Transformer architectures, it can be challenging to create one for your checkpoint. As a part of 🤗 Transformers core philosophy to make the library easy, simple and flexible to use, an `AutoClass` automatically infer and load the correct architecture from a given checkpoint. The `from_pretrained()` method lets you quickly load a pretrained model for any architecture so you don't have to devote time and resources to train a model from scratch. Producing this type of checkpoint-agnostic code means if your code works for one checkpoint, it will work with another checkpoint - as long as it was trained for a similar task - even if the architecture is different.
 
 <Tip>
 
@@ -95,6 +95,14 @@ Easily reuse the same checkpoint to load an architecture for a different task:
 >>> model = AutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
 ```
 
+<Tip warning={true}>
+
+For PyTorch models, the `from_pretrained()` method uses `torch.load()` which internally uses `pickle` and is known to be insecure. In general, never load a model that could have come from an untrusted source, or that could have been tampered with. This security risk is partially mitigated for public models hosted on the Hugging Face Hub, which are [scanned for malware](https://huggingface.co/docs/hub/security-malware) at each commit. See the [Hub documentation](https://huggingface.co/docs/hub/security) for best practices like [signed commit verification](https://huggingface.co/docs/hub/security-gpg#signing-commits-with-gpg) with GPG.
+
+TensorFlow and Flax checkpoints are not affected, and can be loaded within PyTorch architectures using the `from_tf` and `from_flax` kwargs for the `from_pretrained` method to circumvent this issue.
+
+</Tip>
+
 Generally, we recommend using the `AutoTokenizer` class and the `AutoModelFor` class to load pretrained instances of models. This will ensure you load the correct architecture every time. In the next [tutorial](preprocessing), learn how to use your newly loaded tokenizer, feature extractor and processor to preprocess a dataset for fine-tuning.
 </pt>
 <tf>

docs/source/en/big_models.mdx

@@ -114,15 +114,6 @@ If you want to directly load such a sharded checkpoint inside a model without us
 
 ## Low memory loading
 
-Sharded checkpoints reduce the memory usage during step 2 of the worflow mentioned above, but when loadin a pretrained model, why keep the random weights in memory? The option `low_cpu_mem_usage` will destroy the weights of the randomly initialized model, then progressively load the weights inside, then perform a random initialization for potential missing weights (if you are loadding a model with a newly initialized head for a fine-tuning task for instance).
-
-It's very easy to use, just add `low_cpu_mem_usage=True` to your call to [`~PreTrainedModel.from_pretrained`]:
-
-```py
-from transformers import AutoModelForSequenceClas
-
-model = AutoModel.from_pretrained("bert-base-cased", low_cpu_mem_usage=True)
-```
-
-This can be used in conjunction with a sharded checkpoint.
+Sharded checkpoints reduce the memory usage during step 2 of the workflow mentioned above, but in order to use that model in a low memory setting, we recommend leveraging our tools based on the Accelerate library.
 
+Please read the following guide for more information: [Large model loading using Accelerate](./main_classes/model#large-model-loading)

docs/source/en/custom_models.mdx

@@ -289,7 +289,7 @@ from huggingface_hub import notebook_login
 notebook_login()
 ```
 
-You can then push to to your own namespace (or an organization you are a member of) like this:
+You can then push to your own namespace (or an organization you are a member of) like this:
 
 ```py
 resnet50d.push_to_hub("custom-resnet50d")

docs/source/en/glossary.mdx

@@ -44,7 +44,7 @@ specific language governing permissions and limitations under the License.
 Every model is different yet bears similarities with the others. Therefore most models use the same inputs, which are
 detailed here alongside usage examples.
 
-<a id='input-ids'></a>
+
 
 ### Input IDs
 
@@ -113,7 +113,7 @@ we will see
 
 because this is the way a [`BertModel`] is going to expect its inputs.
 
-<a id='attention-mask'></a>
+
 
 ### Attention mask
 
@@ -171,7 +171,7 @@ in the dictionary returned by the tokenizer under the key "attention_mask":
 [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]
 ```
 
-<a id='token-type-ids'></a>
+
 
 ### Token Type IDs
 
@@ -224,7 +224,7 @@ second sequence, corresponding to the "question", has all its tokens represented
 
 Some models, like [`XLNetModel`] use an additional token represented by a `2`.
 
-<a id='position-ids'></a>
+
 
 ### Position IDs
 
@@ -238,7 +238,7 @@ absolute positional embeddings.
 Absolute positional embeddings are selected in the range `[0, config.max_position_embeddings - 1]`. Some models use
 other types of positional embeddings, such as sinusoidal position embeddings or relative position embeddings.
 
-<a id='labels'></a>
+
 
 ### Labels
 
@@ -266,7 +266,7 @@ These labels are different according to the model head, for example:
 The base models (e.g., [`BertModel`]) do not accept labels, as these are the base transformer
 models, simply outputting features.
 
-<a id='decoder-input-ids'></a>
+
 
 ### Decoder input IDs
 
@@ -279,7 +279,6 @@ such models, passing the `labels` is the preferred way to handle training.
 
 Please check each model's docs to see how they handle these input IDs for sequence to sequence training.
 
-<a id='feed-forward-chunking'></a>
 
 ### Feed Forward Chunking
 

docs/source/en/index.mdx

@@ -12,40 +12,38 @@ specific language governing permissions and limitations under the License.
 
 # 🤗 Transformers
 
-State-of-the-art Machine Learning for PyTorch, TensorFlow and JAX.
+State-of-the-art Machine Learning for [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/), and [JAX](https://jax.readthedocs.io/en/latest/).
 
-🤗 Transformers provides APIs to easily download and train state-of-the-art pretrained models. Using pretrained models can reduce your compute costs, carbon footprint, and save you time from training a model from scratch. The models can be used across different modalities such as:
+🤗 Transformers provides APIs and tools to easily download and train state-of-the-art pretrained models. Using pretrained models can reduce your compute costs, carbon footprint, and save you the time and resources required to train a model from scratch. These models support common tasks in different modalities, such as:
 
-* 📝 Text: text classification, information extraction, question answering, summarization, translation, and text generation in over 100 languages.
-* 🖼️ Images: image classification, object detection, and segmentation.
-* 🗣️ Audio: speech recognition and audio classification.
-* 🐙 Multimodal: table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
+📝 **Natural Language Processing**: text classification, named entity recognition, question answering, language modeling, summarization, translation, multiple choice, and text generation.<br>
+🖼️ **Computer Vision**: image classification, object detection, and segmentation.<br>
+🗣️ **Audio**: automatic speech recognition and audio classification.<br>
+🐙 **Multimodal**: table question answering, optical character recognition, information extraction from scanned documents, video classification, and visual question answering.
 
-Our library supports seamless integration between three of the most popular deep learning libraries: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) and [JAX](https://jax.readthedocs.io/en/latest/). Train your model in three lines of code in one framework, and load it for inference with another.
+🤗 Transformers support framework interoperability between PyTorch, TensorFlow, and JAX. This provides the flexibility to use a different framework at each stage of a model's life; train a model in three lines of code in one framework, and load it for inference in another. Models can also be exported to a format like ONNX and TorchScript for deployment in production environments.
 
-Each 🤗 Transformers architecture is defined in a standalone Python module so they can be easily customized for research and experiments.
+Join the growing community on the [Hub](https://huggingface.co/models), [forum](https://discuss.huggingface.co/), or [Discord](https://discord.com/invite/JfAtkvEtRb) today!
 
 ## If you are looking for custom support from the Hugging Face team
 
 <a target="_blank" href="https://huggingface.co/support">
-<img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
+    <img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
 </a><br>
 
 ## Contents
 
-The documentation is organized in five parts:
+The documentation is organized into five sections:
 
-- **GET STARTED** contains a quick tour and installation instructions to get up and running with 🤗 Transformers.
-- **TUTORIALS** are a great place to begin if you are new to our library. This section will help you gain the basic skills you need to start using 🤗 Transformers.
-- **HOW-TO GUIDES** will show you how to achieve a specific goal like fine-tuning a pretrained model for language modeling or how to create a custom model head.
-- **CONCEPTUAL GUIDES** provides more discussion and explanation of the underlying concepts and ideas behind models, tasks, and the design philosophy of 🤗 Transformers. 
-- **API** describes each class and function, grouped in:
+- **GET STARTED** provides a quick tour of the library and installation instructions to get up and running.
+- **TUTORIALS** are a great place to start if you're a beginner. This section will help you gain the basic skills you need to start using the library.
+- **HOW-TO GUIDES** show you how to achieve a specific goal, like finetuning a pretrained model for language modeling or how to write and share a custom model.
+- **CONCEPTUAL GUIDES** offers more discussion and explanation of the underlying concepts and ideas behind models, tasks, and the design philosophy of 🤗 Transformers.
+- **API** describes all classes and functions:
 
-  - **MAIN CLASSES** for the main classes exposing the important APIs of the library.
-  - **MODELS** for the classes and functions related to each model implemented in the library.
-  - **INTERNAL HELPERS** for the classes and functions we use internally.
-
-The library currently contains JAX, PyTorch and TensorFlow implementations, pretrained model weights, usage scripts and conversion utilities for the following models.
+  - **MAIN CLASSES** details the most important classes like configuration, model, tokenizer, and pipeline.
+  - **MODELS** details the classes and functions related to each model implemented in the library.
+  - **INTERNAL HELPERS** details utility classes and functions used internally.
 
 ### Supported models
 
@@ -57,76 +55,95 @@ The library currently contains JAX, PyTorch and TensorFlow implementations, pret
 1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
 1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
 1. **[BERT](model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
-1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
 1. **[BERT For Sequence Generation](model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
-1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
 1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
+1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
 1. **[Blenderbot](model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
 1. **[BlenderbotSmall](model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
+1. **[BLOOM](model_doc/bloom)** (from BigScience workshop) released by the [BigSicence Workshop](https://bigscience.huggingface.co/).
 1. **[BORT](model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
 1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
 1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suárez*, Yoann Dupont, Laurent Romary, Éric Villemonte de la Clergerie, Djamé Seddah and Benoît Sagot.
 1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
-1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
+1. **[CodeGen](model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
 1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
+1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
 1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
 1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
+1. **[CvT](model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
 1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec:  A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
 1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
 1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
-1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
 1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervé Jégou.
 1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
 1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
 1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
+1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
+1. **[Donut](model_doc/donut)** (from NAVER), released together with the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
 1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
 1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by René Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
-1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
 1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
+1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
+1. **[ERNIE](model_doc/ernie)** (from Baidu) released with the paper [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu.
 1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loïc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoît Crabbé, Laurent Besacier, Didier Schwab.
 1. **[FLAVA](model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
 1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
 1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
 1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
 1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
+1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GPT NeoX](model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
 1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
 1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
-1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
+1. **[GroupViT](model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
 1. **[Hubert](model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
 1. **[I-BERT](model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
 1. **[ImageGPT](model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
 1. **[LayoutLM](model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
 1. **[LayoutLMv2](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
-1. **[LayoutXLM](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
+1. **[LayoutLMv3](model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+1. **[LayoutXLM](model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
 1. **[LED](model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LeViT](model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze.
 1. **[Longformer](model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
+1. **[LongT5](model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
 1. **[LUKE](model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
-1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
+1. **[M-CTC-T](model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
 1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
 1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jörg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
 1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
-1. **[MBart](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
-1. **[MBart-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
+1. **[mBART](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
+1. **[mBART-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
 1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
+1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
+1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
 1. **[MT5](model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
+1. **[MVP](model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+1. **[Nezha](model_doc/nezha)** (from Huawei Noah’s Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
+1. **[NLLB](model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
 1. **[Nyströmformer](model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
 1. **[OPT](master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
+1. **[OWL-ViT](model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
 1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
+1. **[PEGASUS-X](model_doc/pegasus_x)** (from Google) released with the paper [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) by Jason Phang, Yao Zhao, and Peter J. Liu.
 1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hénaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, João Carreira.
 1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
 1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
 1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
 1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
+1. **[RAG](model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, Douwe Kiela.
 1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
 1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, Łukasz Kaiser, Anselm Levskaya.
-1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
 1. **[RegNet](model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollár.
+1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Févry, Henry Tsai, M. Johnson, Sebastian Ruder.
 1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
 1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
 1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
@@ -136,32 +153,38 @@ The library currently contains JAX, PyTorch and TensorFlow implementations, pret
 1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
 1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
 1. **[Splinter](model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
-1. **[SqueezeBert](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
+1. **[SqueezeBERT](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
 1. **[Swin Transformer](model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
+1. **[Swin Transformer V2](model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
 1. **[T5](model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[T5v1.1](model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
 1. **[TAPAS](model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweł Krzysztof Nowak, Thomas Müller, Francesco Piccinno and Julian Martin Eisenschlos.
 1. **[TAPEX](model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
+1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
 1. **[Transformer-XL](model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
 1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
+1. **[UL2](model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
 1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
 1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
 1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
+1. **[VideoMAE](model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
 1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
 1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
-1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
-1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, Ross Girshick.
 1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
+1. **[Wav2Vec2-Conformer](model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
 1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
+1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
+1. **[X-CLIP](model_doc/xclip)** (from Microsoft Research) released with the paper [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) by Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling.
 1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
 1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
 1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
 1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
 1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
 1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [​XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
-1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
 1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
+1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
 1. **[YOLOS](model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
 1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
 
@@ -182,73 +205,89 @@ Flax), PyTorch, and/or TensorFlow.
 |            BERT             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |       Bert Generation       |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           BigBird           |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
-|       BigBirdPegasus        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|       BigBird-Pegasus       |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         Blenderbot          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |       BlenderbotSmall       |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|            BLOOM            |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          CamemBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           Canine            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           CANINE            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            CLIP             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|           CodeGen           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          ConvBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|          ConvNext           |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|          ConvNeXT           |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            CTRL             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|             CvT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |        Data2VecAudio        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |        Data2VecText         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |       Data2VecVision        |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |           DeBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |         DeBERTa-v2          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |    Decision Transformer     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            DeiT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            DeiT             |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            DETR             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         DistilBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|          DonutSwin          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             DPR             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |             DPT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           ELECTRA           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |       Encoder decoder       |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|            ERNIE            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 | FairSeq Machine-Translation |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          FlauBERT           |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            Flava            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            FLAVA            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            FNet             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |     Funnel Transformer      |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |            GLPN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           GPT Neo           |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
+|          GPT NeoX           |       ❌       |       ✅       |       ✅        |         ❌         |      ❌      |
 |            GPT-J            |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|          GroupViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           Hubert            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |           I-BERT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          ImageGPT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          LayoutLM           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |         LayoutLMv2          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|         LayoutLMv3          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |             LED             |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|            LeViT            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         Longformer          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           LongT5            |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
 |            LUKE             |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           LXMERT            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|           M-CTC-T           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           M2M100            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           Marian            |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
 |         MaskFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            mBART            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|        MegatronBert         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Megatron-BERT        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         MobileBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|          MobileViT          |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            MPNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|             mT5             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|        Nystromformer        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             MT5             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|             MVP             |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|            Nezha            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|        Nyströmformer        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         OpenAI GPT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |        OpenAI GPT-2         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|             OPT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|             OPT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
+|           OWL-ViT           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           Pegasus           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|          PEGASUS-X          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          Perceiver          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           PLBart            |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         PoolFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         ProphetNet          |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |           QDQBert           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             RAG             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            Realm            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
+|            REALM            |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |          Reformer           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|           RegNet            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           RegNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |           RemBERT           |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
-|           ResNet            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|           ResNet            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |          RetriBERT          |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |           RoBERTa           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |          RoFormer           |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
-|          SegFormer          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          SegFormer          |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |             SEW             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            SEW-D            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |   Speech Encoder decoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
@@ -256,28 +295,32 @@ Flax), PyTorch, and/or TensorFlow.
 |        Speech2Text2         |       ✅       |       ❌       |       ❌        |         ❌         |      ❌      |
 |          Splinter           |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
 |         SqueezeBERT         |       ✅       |       ✅       |       ✅        |         ❌         |      ❌      |
-|            Swin             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|      Swin Transformer       |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
+|     Swin Transformer V2     |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             T5              |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |            TAPAS            |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
-|            TAPEX            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
+|   Trajectory Transformer    |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |       Transformer-XL        |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            TrOCR            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          UniSpeech          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |        UniSpeechSat         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             VAN             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|          VideoMAE           |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            ViLT             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |   Vision Encoder decoder    |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
 |    VisionTextDualEncoder    |       ❌       |       ❌       |       ✅        |         ❌         |      ✅      |
-|         VisualBert          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|         VisualBERT          |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |             ViT             |       ❌       |       ❌       |       ✅        |         ✅         |      ✅      |
 |           ViTMAE            |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |          Wav2Vec2           |       ✅       |       ❌       |       ✅        |         ✅         |      ✅      |
+|     Wav2Vec2-Conformer      |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            WavLM            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|            XGLM             |       ✅       |       ✅       |       ✅        |         ❌         |      ✅      |
+|           X-CLIP            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
+|            XGLM             |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |             XLM             |       ✅       |       ❌       |       ✅        |         ✅         |      ❌      |
+|       XLM-ProphetNet        |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         XLM-RoBERTa         |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |       XLM-RoBERTa-XL        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
-|        XLMProphetNet        |       ✅       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            XLNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
 |            YOLOS            |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |            YOSO             |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |

docs/source/en/installation.mdx

@@ -34,11 +34,16 @@ Start by creating a virtual environment in your project directory:
 python -m venv .env
 ```
 
-Activate the virtual environment:
+Activate the virtual environment. On Linux and MacOs:
 
 ```bash
 source .env/bin/activate
 ```
+Activate Virtual environment on Windows
+
+```bash
+.env/Scripts/activate
+```
 
 Now you're ready to install 🤗 Transformers with the following command:
 
@@ -134,11 +139,11 @@ conda install -c huggingface transformers
 
 ## Cache setup
 
-Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/transformers/`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\transformers`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
+Pretrained models are downloaded and locally cached at: `~/.cache/huggingface/hub`. This is the default directory given by the shell environment variable `TRANSFORMERS_CACHE`. On Windows, the default directory is given by `C:\Users\username\.cache\huggingface\hub`. You can change the shell environment variables shown below - in order of priority - to specify a different cache directory:
 
-1. Shell environment variable (default): `TRANSFORMERS_CACHE`.
-2. Shell environment variable: `HF_HOME` + `transformers/`.
-3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface/transformers`.
+1. Shell environment variable (default): `HUGGINGFACE_HUB_CACHE` or `TRANSFORMERS_CACHE`.
+2. Shell environment variable: `HF_HOME`.
+3. Shell environment variable: `XDG_CACHE_HOME` + `/huggingface`.
 
 <Tip>
 

docs/source/en/internal/generation_utils.mdx

@@ -127,6 +127,9 @@ generation.
 [[autodoc]] TopKLogitsWarper
     - __call__
 
+[[autodoc]] TypicalLogitsWarper
+    - __call__
+
 [[autodoc]] NoRepeatNGramLogitsProcessor
     - __call__
 

docs/source/en/main_classes/callback.mdx

@@ -32,6 +32,7 @@ By default a [`Trainer`] will use the following callbacks:
 - [`~integrations.WandbCallback`] if [wandb](https://www.wandb.com/) is installed.
 - [`~integrations.CometCallback`] if [comet_ml](https://www.comet.ml/site/) is installed.
 - [`~integrations.MLflowCallback`] if [mlflow](https://www.mlflow.org/) is installed.
+- [`~integrations.NeptuneCallback`] if [neptune](https://neptune.ai/) is installed.
 - [`~integrations.AzureMLCallback`] if [azureml-sdk](https://pypi.org/project/azureml-sdk/) is
   installed.
 - [`~integrations.CodeCarbonCallback`] if [codecarbon](https://pypi.org/project/codecarbon/) is
@@ -70,6 +71,8 @@ Here is the list of the available [`TrainerCallback`] in the library:
 
 [[autodoc]] integrations.CodeCarbonCallback
 
+[[autodoc]] integrations.NeptuneCallback
+
 ## TrainerCallback
 
 [[autodoc]] TrainerCallback

docs/source/en/main_classes/deepspeed.mdx

@@ -37,7 +37,7 @@ won't be possible on a single GPU.
 2. If you don't use [`Trainer`] and want to use your own Trainer where you integrated DeepSpeed
    yourself, core functionality functions like `from_pretrained` and `from_config` include integration of essential
    parts of DeepSpeed like `zero.Init` for ZeRO stage 3 and higher. To tap into this feature read the docs on
-   [deepspeed-non-trainer-integration](#deepspeed-non-trainer-integration).
+   [non-Trainer DeepSpeed Integration](#nontrainer-deepspeed-integration).
 
 What is integrated:
 
@@ -1849,7 +1849,6 @@ In this case you usually need to raise the value of `initial_scale_power`. Setti
 
 
 
-<a id='deepspeed-non-trainer-integration'></a>
 
 ## Non-Trainer Deepspeed Integration
 

docs/source/en/main_classes/feature_extractor.mdx

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Feature Extractor
 
-A feature extractor is in charge of preparing input features for a multi-modal model. This includes feature extraction
+A feature extractor is in charge of preparing input features for audio or vision models. This includes feature extraction
 from sequences, *e.g.*, pre-processing audio files to Log-Mel Spectrogram features, feature extraction from images
 *e.g.* cropping image image files, but also padding, normalization, and conversion to Numpy, PyTorch, and TensorFlow
 tensors.

docs/source/en/main_classes/logging.mdx

@@ -40,29 +40,17 @@ Additionally, some `warnings` can be disabled by setting the environment variabl
 TRANSFORMERS_NO_ADVISORY_WARNINGS=1 ./myprogram.py
 ```
 
-Here is an example of how to use `logging` in a module:
+Here is an example of how to use the same logger as the library in your own module or script:
 
 ```python
 from transformers.utils import logging
 
 logging.set_verbosity_info()
-logger = logging.get_logger(__name__)
+logger = logging.get_logger("transformers")
 logger.info("INFO")
 logger.warning("WARN")
 ```
 
-Above, a `logger` instance is created from `logging.get_logger(__name__)`. If you want to use `logging` in a script, you shouldn't pass `__name__` to `logging.get_logger`. For example:
-
-```python
-from transformers.utils import logging
-
-if __name__ == "__main__":
-    logging.set_verbosity_info()
-    # leave it empy or use a string
-    logger = logging.get_logger()
-    logger.info("INFO")
-    logger.warning("WARN")
-```
 
 All the methods of this logging module are documented below, the main ones are
 [`logging.get_verbosity`] to get the current level of verbosity in the logger and

docs/source/en/main_classes/model.mdx

@@ -38,6 +38,75 @@ for text generation, [`~generation_utils.GenerationMixin`] (for the PyTorch mode
 
 <a id='from_pretrained-torch-dtype'></a>
 
+### Large model loading
+
+In Transformers 4.20.0, the [`~PreTrainedModel.from_pretrained`] method has been reworked to accommodate large models using [Accelerate](https://huggingface.co/docs/accelerate/big_modeling). This requires Accelerate >= 0.9.0 and PyTorch >= 1.9.0. Instead of creating the full model, then loading the pretrained weights inside it (which takes twice the size of the model in RAM, one for the randomly initialized model, one for the weights), there is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded.
+
+This option can be activated with `low_cpu_mem_usage=True`. The model is first created on the Meta device (with empty weights) and the state dict is then loaded inside it (shard by shard in the case of a sharded checkpoint). This way the maximum RAM used is the full size of the model only.
+
+```py
+from transformers import AutoModelForSeq2SeqLM
+
+t0pp = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0pp", low_cpu_mem_usage=True)
+```
+
+Moreover, you can directly place the model on different devices if it doesn't fully fit in RAM (only works for inference for now). With `device_map="auto"`, Accelerate will determine where to put each layer to maximize the use of your fastest devices (GPUs) and offload the rest on the CPU, or even the hard drive if you don't have enough GPU RAM (or CPU RAM). Even if the model is split across several devices, it will run as you would normally expect.
+
+When passing a `device_map`, `low_cpu_mem_usage` is automatically set to `True`, so you don't need to specify it:
+
+```py
+from transformers import AutoModelForSeq2SeqLM
+
+t0pp = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0pp", device_map="auto")
+```
+
+You can inspect how the model was split across devices by looking at its `hf_device_map` attribute:
+
+```py
+t0pp.hf_device_map
+```
+
+```python out
+{'shared': 0,
+ 'decoder.embed_tokens': 0,
+ 'encoder': 0,
+ 'decoder.block.0': 0,
+ 'decoder.block.1': 1,
+ 'decoder.block.2': 1,
+ 'decoder.block.3': 1,
+ 'decoder.block.4': 1,
+ 'decoder.block.5': 1,
+ 'decoder.block.6': 1,
+ 'decoder.block.7': 1,
+ 'decoder.block.8': 1,
+ 'decoder.block.9': 1,
+ 'decoder.block.10': 1,
+ 'decoder.block.11': 1,
+ 'decoder.block.12': 1,
+ 'decoder.block.13': 1,
+ 'decoder.block.14': 1,
+ 'decoder.block.15': 1,
+ 'decoder.block.16': 1,
+ 'decoder.block.17': 1,
+ 'decoder.block.18': 1,
+ 'decoder.block.19': 1,
+ 'decoder.block.20': 1,
+ 'decoder.block.21': 1,
+ 'decoder.block.22': 'cpu',
+ 'decoder.block.23': 'cpu',
+ 'decoder.final_layer_norm': 'cpu',
+ 'decoder.dropout': 'cpu',
+ 'lm_head': 'cpu'}
+```
+
+You can also write your own device map following the same format (a dictionary layer name to device). It should map all parameters of the model to a given device, but you don't have to detail where all the submosules of one layer go if that layer is entirely on the same device. For instance, the following device map would work properly for T0pp (as long as you have the GPU memory):
+
+```python
+device_map = {"shared": 0, "encoder": 0, "decoder": 1, "lm_head": 1}
+```
+
+Another way to minimize the memory impact of your model is to instantiate it at a lower precision dtype (like `torch.float16`) or use direct quantization techniques as described below.
+
 ### Model Instantiation dtype
 
 Under Pytorch a model normally gets instantiated with `torch.float32` format. This can be an issue if one tries to
@@ -65,7 +134,6 @@ model = AutoModel.from_config(config)
 Due to Pytorch design, this functionality is only available for floating dtypes.
 
 
-
 ## ModuleUtilsMixin
 
 [[autodoc]] modeling_utils.ModuleUtilsMixin

docs/source/en/main_classes/output.mdx

@@ -16,7 +16,7 @@ All models have outputs that are instances of subclasses of [`~utils.ModelOutput
 data structures containing all the information returned by the model, but that can also be used as tuples or
 dictionaries.
 
-Let's see of this looks on an example:
+Let's see how this looks in an example:
 
 ```python
 from transformers import BertTokenizer, BertForSequenceClassification
@@ -136,6 +136,30 @@ documented on their corresponding model page.
 
 [[autodoc]] modeling_outputs.Seq2SeqQuestionAnsweringModelOutput
 
+## SemanticSegmenterOutput
+
+[[autodoc]] modeling_outputs.SemanticSegmenterOutput
+
+## ImageClassifierOutput
+
+[[autodoc]] modeling_outputs.ImageClassifierOutput
+
+## ImageClassifierOutputWithNoAttention
+
+[[autodoc]] modeling_outputs.ImageClassifierOutputWithNoAttention
+
+## DepthEstimatorOutput
+
+[[autodoc]] modeling_outputs.DepthEstimatorOutput
+
+## Wav2Vec2BaseModelOutput
+
+[[autodoc]] modeling_outputs.Wav2Vec2BaseModelOutput
+
+## XVectorOutput
+
+[[autodoc]] modeling_outputs.XVectorOutput
+
 ## TFBaseModelOutput
 
 [[autodoc]] modeling_tf_outputs.TFBaseModelOutput

docs/source/en/main_classes/pipelines.mdx

@@ -25,10 +25,12 @@ There are two categories of pipeline abstractions to be aware about:
   - [`AudioClassificationPipeline`]
   - [`AutomaticSpeechRecognitionPipeline`]
   - [`ConversationalPipeline`]
+  - [`DocumentQuestionAnsweringPipeline`]
   - [`FeatureExtractionPipeline`]
   - [`FillMaskPipeline`]
   - [`ImageClassificationPipeline`]
   - [`ImageSegmentationPipeline`]
+  - [`ImageToTextPipeline`]
   - [`ObjectDetectionPipeline`]
   - [`QuestionAnsweringPipeline`]
   - [`SummarizationPipeline`]
@@ -38,6 +40,7 @@ There are two categories of pipeline abstractions to be aware about:
   - [`Text2TextGenerationPipeline`]
   - [`TokenClassificationPipeline`]
   - [`TranslationPipeline`]
+  - [`VisualQuestionAnsweringPipeline`]
   - [`ZeroShotClassificationPipeline`]
   - [`ZeroShotImageClassificationPipeline`]
 
@@ -340,6 +343,12 @@ That should enable you to do all the custom code you want.
     - __call__
     - all
 
+### DocumentQuestionAnsweringPipeline
+
+[[autodoc]] DocumentQuestionAnsweringPipeline
+    - __call__
+    - all
+
 ### FeatureExtractionPipeline
 
 [[autodoc]] FeatureExtractionPipeline
@@ -364,6 +373,12 @@ That should enable you to do all the custom code you want.
     - __call__
     - all
 
+### ImageToTextPipeline
+
+[[autodoc]] ImageToTextPipeline
+    - __call__
+    - all
+
 ### NerPipeline
 
 [[autodoc]] NerPipeline
@@ -423,6 +438,12 @@ See [`TokenClassificationPipeline`] for all details.
     - __call__
     - all
 
+### VisualQuestionAnsweringPipeline
+
+[[autodoc]] VisualQuestionAnsweringPipeline
+    - __call__
+    - all
+
 ### ZeroShotClassificationPipeline
 
 [[autodoc]] ZeroShotClassificationPipeline

docs/source/en/main_classes/trainer.mdx

@@ -291,10 +291,10 @@ Also if you do set this environment variable it's the best to set it in your `~/
 The [`Trainer`] has been extended to support libraries that may dramatically improve your training
 time and fit much bigger models.
 
-Currently it supports third party solutions, [DeepSpeed](https://github.com/microsoft/DeepSpeed) and [FairScale](https://github.com/facebookresearch/fairscale/), which implement parts of the paper [ZeRO: Memory Optimizations
+Currently it supports third party solutions, [DeepSpeed](https://github.com/microsoft/DeepSpeed), [PyTorch FSDP](https://pytorch.org/docs/stable/fsdp.html) and [FairScale](https://github.com/facebookresearch/fairscale/), which implement parts of the paper [ZeRO: Memory Optimizations
 Toward Training Trillion Parameter Models, by Samyam Rajbhandari, Jeff Rasley, Olatunji Ruwase, Yuxiong He](https://arxiv.org/abs/1910.02054).
 
-This provided support is new and experimental as of this writing.
+This provided support is new and experimental as of this writing. While the support for DeepSpeed and PyTorch FSDP is active and we welcome issues around it, we don't support the FairScale integration anymore since it has been integrated in PyTorch main (see the [PyTorch FSDP integration](#pytorch-fully-sharded-data-parallel))
 
 <a id='zero-install-notes'></a>
 
@@ -408,6 +408,12 @@ As always make sure to edit the paths in the example to match your situation.
 
 ### FairScale
 
+<Tip warning={true}>
+
+This integration is not supported anymore, we recommend you either use DeepSpeed or PyTorch FSDP.
+
+</Tip>
+
 By integrating [FairScale](https://github.com/facebookresearch/fairscale/) the [`Trainer`]
 provides support for the following features from [the ZeRO paper](https://arxiv.org/abs/1910.02054):
 
@@ -561,14 +567,22 @@ as the model saving with FSDP activated is only available with recent fixes.
   For this, add `--fsdp full_shard` to the command line arguments. 
   - SHARD_GRAD_OP : Shards optimizer states + gradients across data parallel workers/GPUs.
     For this, add `--fsdp shard_grad_op` to the command line arguments.
+  - NO_SHARD : No sharding. For this, add `--fsdp no_shard` to the command line arguments.
 - To offload the parameters and gradients to the CPU, 
 add `--fsdp "full_shard offload"` or `--fsdp "shard_grad_op offload"` to the command line arguments.
 -  To automatically recursively wrap layers with FSDP using `default_auto_wrap_policy`, 
 add `--fsdp "full_shard auto_wrap"` or `--fsdp "shard_grad_op auto_wrap"` to the command line arguments.
 - To enable both CPU offloading and auto wrapping, 
 add `--fsdp "full_shard offload auto_wrap"` or `--fsdp "shard_grad_op offload auto_wrap"` to the command line arguments.
-- If auto wrapping is enabled, please add `--fsdp_min_num_params <number>` to command line arguments.
-It specifies FSDP's minimum number of parameters for Default Auto Wrapping.
+- If auto wrapping is enabled, you can either use transformer based auto wrap policy or size based auto wrap policy.
+  - For transformer based auto wrap policy, please add `--fsdp_transformer_layer_cls_to_wrap <value>` to command line arguments.
+  This specifies the transformer layer class name (case-sensitive) to wrap ,e.g, `BertLayer`, `GPTJBlock`, `T5Block` ....
+  This is important because submodules that share weights (e.g., embedding layer) should not end up in different FSDP wrapped units. 
+  Using this policy, wrapping happens for each block containing Multi-Head Attention followed by couple of MLP layers. 
+  Remaining layers including the shared embeddings are conviniently wrapped in same outermost FSDP unit.
+  Therefore, use this for transformer based models.
+  - For size based auto wrap policy, please add `--fsdp_min_num_params <number>` to command line arguments.
+  It specifies FSDP's minimum number of parameters for auto wrapping.
 
 **Few caveats to be aware of**
 - Mixed precision is currently not supported with FSDP as we wait for PyTorch to fix support for it.
@@ -577,6 +591,66 @@ More details in this [issues](https://github.com/pytorch/pytorch/issues/75676).
 More details mentioned in this [issue](https://github.com/pytorch/pytorch/issues/76501)
 (`The original model parameters' .grads are not set, meaning that they cannot be optimized separately (which is why we cannot support multiple parameter groups)`).
 
+### Using Trainer for accelerated PyTorch Training on Mac 
+
+With PyTorch v1.12 release, developers and researchers can take advantage of Apple silicon GPUs for significantly faster model training. 
+This unlocks the ability to perform machine learning workflows like prototyping and fine-tuning locally, right on Mac.
+Apple's Metal Performance Shaders (MPS) as a backend for PyTorch enables this and can be used via the new `"mps"` device. 
+This will map computational graphs and primitives on the MPS Graph framework and tuned kernels provided by MPS.
+For more information please refer official documents [Introducing Accelerated PyTorch Training on Mac](https://pytorch.org/blog/introducing-accelerated-pytorch-training-on-mac/)
+and [MPS BACKEND](https://pytorch.org/docs/stable/notes/mps.html). 
+
+<Tip warning={false}>
+
+We strongly recommend to install PyTorch >= 1.13 (nightly version at the time of writing) on your MacOS machine. 
+It has major fixes related to model correctness and performance improvements for transformer based models.
+Please refer to https://github.com/pytorch/pytorch/issues/82707 for more details.
+
+</Tip>
+
+**Benefits of Training and Inference using Apple Silicon Chips**
+
+1. Enables users to train larger networks or batch sizes locally
+2. Reduces data retrieval latency and provides the GPU with direct access to the full memory store due to unified memory architecture. 
+Therefore, improving end-to-end performance.
+3. Reduces costs associated with cloud-based development or the need for additional local GPUs.
+
+**Pre-requisites**: To install torch with mps support, 
+please follow this nice medium article [GPU-Acceleration Comes to PyTorch on M1 Macs](https://medium.com/towards-data-science/gpu-acceleration-comes-to-pytorch-on-m1-macs-195c399efcc1).
+
+**Usage**:
+User has to just pass `--use_mps_device` argument. 
+For example, you can run the offical Glue text classififcation task (from the root folder) using Apple Silicon GPU with below command:
+
+```bash
+export TASK_NAME=mrpc
+
+python examples/pytorch/text-classification/run_glue.py \
+  --model_name_or_path bert-base-cased \
+  --task_name $TASK_NAME \
+  --do_train \
+  --do_eval \
+  --max_seq_length 128 \
+  --per_device_train_batch_size 32 \
+  --learning_rate 2e-5 \
+  --num_train_epochs 3 \
+  --output_dir /tmp/$TASK_NAME/ \
+  --use_mps_device \
+  --overwrite_output_dir
+```
+
+**A few caveats to be aware of**
+
+1. Some PyTorch operations have not been implemented in mps and will throw an error. 
+One way to get around that is to set the environment variable `PYTORCH_ENABLE_MPS_FALLBACK=1`, 
+which will fallback to CPU for these operations. It still throws a UserWarning however.
+2. Distributed setups `gloo` and `nccl` are not working with `mps` device. 
+This means that currently only single GPU of `mps` device type can be used.
+
+Finally, please, remember that, 🤗 `Trainer` only integrates MPS backend, therefore if you
+have any problems or questions with regards to MPS backend usage, please, 
+file an issue with [PyTorch GitHub](https://github.com/pytorch/pytorch/issues).
+
 Sections that were moved:
 
 [ <a href="./deepspeed#deepspeed-trainer-integration">DeepSpeed</a><a id="deepspeed"></a>

docs/source/en/model_doc/auto.mdx

@@ -114,14 +114,26 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] AutoModelForTableQuestionAnswering
 
+## AutoModelForDocumentQuestionAnswering
+
+[[autodoc]] AutoModelForDocumentQuestionAnswering
+
 ## AutoModelForImageClassification
 
 [[autodoc]] AutoModelForImageClassification
 
+## AutoModelForVideoClassification
+
+[[autodoc]] AutoModelForVideoClassification
+
 ## AutoModelForVision2Seq
 
 [[autodoc]] AutoModelForVision2Seq
 
+## AutoModelForVisualQuestionAnswering
+
+[[autodoc]] AutoModelForVisualQuestionAnswering
+
 ## AutoModelForAudioClassification
 
 [[autodoc]] AutoModelForAudioClassification
@@ -178,6 +190,10 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] TFAutoModelForImageClassification
 
+## TFAutoModelForSemanticSegmentation
+
+[[autodoc]] TFAutoModelForSemanticSegmentation
+
 ## TFAutoModelForMaskedLM
 
 [[autodoc]] TFAutoModelForMaskedLM
@@ -202,6 +218,10 @@ Likewise, if your `NewModel` is a subclass of [`PreTrainedModel`], make sure its
 
 [[autodoc]] TFAutoModelForTableQuestionAnswering
 
+## TFAutoModelForDocumentQuestionAnswering
+
+[[autodoc]] TFAutoModelForDocumentQuestionAnswering
+
 ## TFAutoModelForTokenClassification
 
 [[autodoc]] TFAutoModelForTokenClassification

docs/source/en/model_doc/bert.mdx

@@ -58,6 +58,10 @@ This model was contributed by [thomwolf](https://huggingface.co/thomwolf). The o
 
 [[autodoc]] BertTokenizerFast
 
+## TFBertTokenizer
+
+[[autodoc]] TFBertTokenizer
+
 ## Bert specific outputs
 
 [[autodoc]] models.bert.modeling_bert.BertForPreTrainingOutput

docs/source/en/model_doc/bloom.mdx

@@ -0,0 +1,57 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# BLOOM
+
+## Overview
+
+The BLOOM model has been proposed with its various versions through the [BigScience Workshop](https://bigscience.huggingface.co/). BigScience is inspired by other open science initiatives where researchers have pooled their time and resources to collectively achieve a higher impact.
+The architecture of BLOOM is essentially similar to GPT3 (auto-regressive model for next token prediction), but has been trained on 46 different languages and 13 programming languages.
+Several smaller versions of the models have been trained on the same dataset. BLOOM is available in the following versions:
+
+- [bloom-560m](https://huggingface.co/bigscience/bloom-560m)
+- [bloom-1b1](https://huggingface.co/bigscience/bloom-1b1)
+- [bloom-1b7](https://huggingface.co/bigscience/bloom-1b7)
+- [bloom-3b](https://huggingface.co/bigscience/bloom-3b)
+- [bloom-7b1](https://huggingface.co/bigscience/bloom-7b1)
+- [bloom](https://huggingface.co/bigscience/bloom) (176B parameters)
+
+
+## BloomConfig
+
+[[autodoc]] BloomConfig
+    - all
+
+## BloomModel
+
+[[autodoc]] BloomModel
+    - forward
+
+## BloomTokenizerFast
+
+[[autodoc]] BloomTokenizerFast
+    - all
+
+## BloomForCausalLM
+
+[[autodoc]] BloomForCausalLM
+    - forward
+
+## BloomForSequenceClassification
+
+[[autodoc]] BloomForSequenceClassification
+    - forward
+
+## BloomForTokenClassification
+
+[[autodoc]] BloomForTokenClassification
+    - forward

docs/source/en/model_doc/codegen.mdx

@@ -0,0 +1,81 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# CodeGen
+
+## Overview
+
+The CodeGen model was proposed in [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, and Caiming Xiong.
+
+CodeGen is an autoregressive language model for program synthesis trained sequentially on [The Pile](https://pile.eleuther.ai/), BigQuery, and BigPython.
+
+The abstract from the paper is the following:
+
+*Program synthesis strives to generate a computer program as a solution to a given problem specification. We propose a conversational program synthesis approach via large language models, which addresses the challenges of searching over a vast program space and user intent specification faced in prior approaches. Our new approach casts the process of writing a specification and program as a multi-turn conversation between a user and a system. It treats program synthesis as a sequence prediction problem, in which the specification is expressed in natural language and the desired program is conditionally sampled. We train a family of large language models, called CodeGen, on natural language and programming language data. With weak supervision in the data and the scaling up of data size and model size, conversational capacities emerge from the simple autoregressive language modeling. To study the model behavior on conversational program synthesis, we develop a multi-turn programming benchmark (MTPB), where solving each problem requires multi-step synthesis via multi-turn conversation between the user and the model. Our findings show the emergence of conversational capabilities and the effectiveness of the proposed conversational program synthesis paradigm. In addition, our model CodeGen (with up to 16B parameters trained on TPU-v4) outperforms OpenAI's Codex on the HumanEval benchmark. We make the training library JaxFormer including checkpoints available as open source contribution: [this https URL](https://github.com/salesforce/codegen).* 
+
+This model was contributed by [Hiroaki Hayashi](https://huggingface.co/rooa).
+The original code can be found [here](https://github.com/salesforce/codegen).
+
+## Checkpoint Naming
+
+* CodeGen model [checkpoints](https://huggingface.co/models?other=codegen) are available on different pre-training data with variable sizes.
+* The format is: `Salesforce/codegen-{size}-{data}`, where
+  * `size`: `350M`, `2B`, `6B`, `16B`
+  * `data`: 
+    * `nl`: Pre-trained on the Pile
+    * `multi`: Initialized with `nl`, then further pre-trained on multiple programming languages data
+    * `mono`: Initialized with `multi`, then further pre-trained on Python data
+* For example, `Salesforce/codegen-350M-mono` offers a 350 million-parameter checkpoint pre-trained sequentially on the Pile, multiple programming languages, and Python.
+
+## How to use
+
+```python
+>>> from transformers import AutoModelForCausalLM, AutoTokenizer
+
+>>> checkpoint = "Salesforce/codegen-350M-mono"
+>>> model = AutoModelForCausalLM.from_pretrained(checkpoint)
+>>> tokenizer = AutoTokenizer.from_pretrained(checkpoint)
+
+>>> text = "def hello_world():"
+
+>>> completion = model.generate(**tokenizer(text, return_tensors="pt"))
+
+>>> print(tokenizer.decode(completion[0]))
+def hello_world():
+    print("Hello World")
+
+hello_world()
+```
+
+## CodeGenConfig
+
+[[autodoc]] CodeGenConfig
+    - all
+
+## CodeGenTokenizer
+
+[[autodoc]] CodeGenTokenizer
+    - save_vocabulary
+
+## CodeGenTokenizerFast
+
+[[autodoc]] CodeGenTokenizerFast
+
+## CodeGenModel
+
+[[autodoc]] CodeGenModel
+    - forward
+
+## CodeGenForCausalLM
+
+[[autodoc]] CodeGenForCausalLM
+    - forward

docs/source/en/model_doc/cvt.mdx

@@ -0,0 +1,53 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# Convolutional Vision Transformer (CvT)
+
+## Overview
+
+The CvT model was proposed in [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan and Lei Zhang. The Convolutional vision Transformer (CvT) improves the [Vision Transformer (ViT)](vit) in performance and efficiency by introducing convolutions into ViT to yield the best of both designs.
+
+The abstract from the paper is the following:
+
+*We present in this paper a new architecture, named Convolutional vision Transformer (CvT), that improves Vision Transformer (ViT) 
+in performance and efficiency by introducing convolutions into ViT to yield the best of both designs. This is accomplished through 
+two primary modifications: a hierarchy of Transformers containing a new convolutional token embedding, and a convolutional Transformer 
+block leveraging a convolutional projection. These changes introduce desirable properties of convolutional neural networks (CNNs) 
+to the ViT architecture (\ie shift, scale, and distortion invariance) while maintaining the merits of Transformers (\ie dynamic attention, 
+global context, and better generalization). We validate CvT by conducting extensive experiments, showing that this approach achieves 
+state-of-the-art performance over other Vision Transformers and ResNets on ImageNet-1k, with fewer parameters and lower FLOPs. In addition, 
+performance gains are maintained when pretrained on larger datasets (\eg ImageNet-22k) and fine-tuned to downstream tasks. Pre-trained on 
+ImageNet-22k, our CvT-W24 obtains a top-1 accuracy of 87.7\% on the ImageNet-1k val set. Finally, our results show that the positional encoding, 
+a crucial component in existing Vision Transformers, can be safely removed in our model, simplifying the design for higher resolution vision tasks.*
+
+Tips:
+
+- CvT models are regular Vision Transformers, but trained with convolutions. They outperform the [original model (ViT)](vit) when fine-tuned on ImageNet-1K and CIFAR-100.
+- You can check out demo notebooks regarding inference as well as fine-tuning on custom data [here](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/VisionTransformer) (you can just replace [`ViTFeatureExtractor`] by [`AutoFeatureExtractor`] and [`ViTForImageClassification`] by [`CvtForImageClassification`]).
+- The available checkpoints are either (1) pre-trained on [ImageNet-22k](http://www.image-net.org/) (a collection of 14 million images and 22k classes) only, (2) also fine-tuned on ImageNet-22k or (3) also fine-tuned on [ImageNet-1k](http://www.image-net.org/challenges/LSVRC/2012/) (also referred to as ILSVRC 2012, a collection of 1.3 million
+  images and 1,000 classes).
+
+This model was contributed by [anugunj](https://huggingface.co/anugunj). The original code can be found [here](https://github.com/microsoft/CvT).
+
+## CvtConfig
+
+[[autodoc]] CvtConfig
+
+## CvtModel
+
+[[autodoc]] CvtModel
+    - forward
+
+## CvtForImageClassification
+
+[[autodoc]] CvtForImageClassification
+    - forward

docs/source/en/model_doc/data2vec.mdx

@@ -37,9 +37,12 @@ Tips:
 - For Data2VecAudio, preprocessing is identical to [`Wav2Vec2Model`], including feature extraction
 - For Data2VecText, preprocessing is identical to [`RobertaModel`], including tokenization.
 - For Data2VecVision, preprocessing is identical to [`BeitModel`], including feature extraction.
+- To know how a pre-trained Data2Vec vision model can be fine-tuned on the task of image classification, you can check out
+[this notebook](https://colab.research.google.com/github/sayakpaul/TF-2.0-Hacks/blob/master/data2vec_vision_image_classification.ipynb).
+
 
 This model was contributed by [edugp](https://huggingface.co/edugp) and [patrickvonplaten](https://huggingface.co/patrickvonplaten).
-[sayakpaul](https://github.com/sayakpaul) contributed Data2Vec for vision in TensorFlow.
+[sayakpaul](https://github.com/sayakpaul) and [Rocketknight1](https://github.com/Rocketknight1) contributed Data2Vec for vision in TensorFlow.
 
 The original code (for NLP and Speech) can be found [here](https://github.com/pytorch/fairseq/tree/main/examples/data2vec).
 The original code for vision can be found [here](https://github.com/facebookresearch/data2vec_vision/tree/main/beit).
@@ -141,4 +144,9 @@ The original code for vision can be found [here](https://github.com/facebookrese
 ## TFData2VecVisionForImageClassification
 
 [[autodoc]] TFData2VecVisionForImageClassification
-    - call
+    - call
+
+## TFData2VecVisionForSemanticSegmentation
+
+[[autodoc]] TFData2VecVisionForSemanticSegmentation
+    - call

docs/source/en/model_doc/deit.mdx

@@ -69,7 +69,7 @@ Tips:
   *facebook/deit-base-patch16-384*. Note that one should use [`DeiTFeatureExtractor`] in order to
   prepare images for the model.
 
-This model was contributed by [nielsr](https://huggingface.co/nielsr).
+This model was contributed by [nielsr](https://huggingface.co/nielsr). The TensorFlow version of this model was added by [amyeroberts](https://huggingface.co/amyeroberts).
 
 
 ## DeiTConfig
@@ -100,3 +100,23 @@ This model was contributed by [nielsr](https://huggingface.co/nielsr).
 
 [[autodoc]] DeiTForImageClassificationWithTeacher
     - forward
+
+## TFDeiTModel
+
+[[autodoc]] TFDeiTModel
+    - call
+
+## TFDeiTForMaskedImageModeling
+
+[[autodoc]] TFDeiTForMaskedImageModeling
+    - call
+
+## TFDeiTForImageClassification
+
+[[autodoc]] TFDeiTForImageClassification
+    - call
+
+## TFDeiTForImageClassificationWithTeacher
+
+[[autodoc]] TFDeiTForImageClassificationWithTeacher
+    - call

docs/source/en/model_doc/detr.mdx

@@ -113,6 +113,28 @@ Tips:
 - The size of the images will determine the amount of memory being used, and will thus determine the `batch_size`.
   It is advised to use a batch size of 2 per GPU. See [this Github thread](https://github.com/facebookresearch/detr/issues/150) for more info.
 
+There are three ways to instantiate a DETR model (depending on what you prefer):
+  
+Option 1: Instantiate DETR with pre-trained weights for entire model
+```py
+>>> from transformers import DetrForObjectDetection
+
+>>> model = DetrForObjectDetection.from_pretrained("facebook/resnet-50")
+```
+
+Option 2: Instantiate DETR with randomly initialized weights for Transformer, but pre-trained weights for backbone
+```py
+>>> from transformers import DetrConfig, DetrForObjectDetection
+
+>>> config = DetrConfig()
+>>> model = DetrForObjectDetection(config)
+```
+Option 3: Instantiate DETR with randomly initialized weights for backbone + Transformer
+```py
+>>> config = DetrConfig(use_pretrained_backbone=False)
+>>> model = DetrForObjectDetection(config)
+```
+
 As a summary, consider the following table:
 
 | Task | Object detection | Instance segmentation | Panoptic segmentation |
@@ -166,4 +188,4 @@ mean Average Precision (mAP) and Panoptic Quality (PQ). The latter objects are i
 ## DetrForSegmentation
 
 [[autodoc]] DetrForSegmentation
-    - forward
+    - forward

docs/source/en/model_doc/donut.mdx

@@ -0,0 +1,214 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the
+License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License. -->
+
+# Donut
+
+## Overview
+
+The Donut model was proposed in [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by
+Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
+Donut consists of an image Transformer encoder and an autoregressive text Transformer decoder to perform document understanding
+tasks such as document image classification, form understanding and visual question answering.
+
+The abstract from the paper is the following:
+
+*Understanding document images (e.g., invoices) is a core but challenging task since it requires complex functions such as reading text and a holistic understanding of the document. Current Visual Document Understanding (VDU) methods outsource the task of reading text to off-the-shelf Optical Character Recognition (OCR) engines and focus on the understanding task with the OCR outputs. Although such OCR-based approaches have shown promising performance, they suffer from 1) high computational costs for using OCR; 2) inflexibility of OCR models on languages or types of document; 3) OCR error propagation to the subsequent process. To address these issues, in this paper, we introduce a novel OCR-free VDU model named Donut, which stands for Document understanding transformer. As the first step in OCR-free VDU research, we propose a simple architecture (i.e., Transformer) with a pre-training objective (i.e., cross-entropy loss). Donut is conceptually simple yet effective. Through extensive experiments and analyses, we show a simple OCR-free VDU model, Donut, achieves state-of-the-art performances on various VDU tasks in terms of both speed and accuracy. In addition, we offer a synthetic data generator that helps the model pre-training to be flexible in various languages and domains.*
+
+<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/model_doc/donut_architecture.jpg"
+alt="drawing" width="600"/> 
+
+<small> Donut high-level overview. Taken from the <a href="https://arxiv.org/abs/2111.15664">original paper</a>. </small>
+
+This model was contributed by [nielsr](https://huggingface.co/nielsr). The original code can be found
+[here](https://github.com/clovaai/donut).
+
+Tips:
+
+- The quickest way to get started with Donut is by checking the [tutorial
+  notebooks](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/Donut), which show how to use the model
+  at inference time as well as fine-tuning on custom data.
+- Donut is always used within the [VisionEncoderDecoder](vision-encoder-decoder) framework.
+
+## Inference
+
+Donut's [`VisionEncoderDecoder`] model accepts images as input and makes use of
+[`~generation_utils.GenerationMixin.generate`] to autoregressively generate text given the input image.
+
+The [`DonutFeatureExtractor`] class is responsible for preprocessing the input image and
+[`XLMRobertaTokenizer`/`XLMRobertaTokenizerFast`] decodes the generated target tokens to the target string. The
+[`DonutProcessor`] wraps [`DonutFeatureExtractor`] and [`XLMRobertaTokenizer`/`XLMRobertaTokenizerFast`]
+into a single instance to both extract the input features and decode the predicted token ids.
+
+- Step-by-step Document Image Classification
+
+```py
+>>> import re
+
+>>> from transformers import DonutProcessor, VisionEncoderDecoderModel
+>>> from datasets import load_dataset
+>>> import torch
+
+>>> processor = DonutProcessor.from_pretrained("naver-clova-ix/donut-base-finetuned-rvlcdip")
+>>> model = VisionEncoderDecoderModel.from_pretrained("naver-clova-ix/donut-base-finetuned-rvlcdip")
+
+>>> device = "cuda" if torch.cuda.is_available() else "cpu"
+>>> model.to(device)  # doctest: +IGNORE_RESULT
+
+>>> # load document image
+>>> dataset = load_dataset("hf-internal-testing/example-documents", split="test")
+>>> image = dataset[1]["image"]
+
+>>> # prepare decoder inputs
+>>> task_prompt = "<s_rvlcdip>"
+>>> decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
+
+>>> pixel_values = processor(image, return_tensors="pt").pixel_values
+
+>>> outputs = model.generate(
+...     pixel_values.to(device),
+...     decoder_input_ids=decoder_input_ids.to(device),
+...     max_length=model.decoder.config.max_position_embeddings,
+...     early_stopping=True,
+...     pad_token_id=processor.tokenizer.pad_token_id,
+...     eos_token_id=processor.tokenizer.eos_token_id,
+...     use_cache=True,
+...     num_beams=1,
+...     bad_words_ids=[[processor.tokenizer.unk_token_id]],
+...     return_dict_in_generate=True,
+... )
+
+>>> sequence = processor.batch_decode(outputs.sequences)[0]
+>>> sequence = sequence.replace(processor.tokenizer.eos_token, "").replace(processor.tokenizer.pad_token, "")
+>>> sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()  # remove first task start token
+>>> print(processor.token2json(sequence))
+{'class': 'advertisement'}
+```
+
+- Step-by-step Document Parsing
+
+```py
+>>> import re
+
+>>> from transformers import DonutProcessor, VisionEncoderDecoderModel
+>>> from datasets import load_dataset
+>>> import torch
+
+>>> processor = DonutProcessor.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
+>>> model = VisionEncoderDecoderModel.from_pretrained("naver-clova-ix/donut-base-finetuned-cord-v2")
+
+>>> device = "cuda" if torch.cuda.is_available() else "cpu"
+>>> model.to(device)  # doctest: +IGNORE_RESULT
+
+>>> # load document image
+>>> dataset = load_dataset("hf-internal-testing/example-documents", split="test")
+>>> image = dataset[2]["image"]
+
+>>> # prepare decoder inputs
+>>> task_prompt = "<s_cord-v2>"
+>>> decoder_input_ids = processor.tokenizer(task_prompt, add_special_tokens=False, return_tensors="pt").input_ids
+
+>>> pixel_values = processor(image, return_tensors="pt").pixel_values
+
+>>> outputs = model.generate(
+...     pixel_values.to(device),
+...     decoder_input_ids=decoder_input_ids.to(device),
+...     max_length=model.decoder.config.max_position_embeddings,
+...     early_stopping=True,
+...     pad_token_id=processor.tokenizer.pad_token_id,
+...     eos_token_id=processor.tokenizer.eos_token_id,
+...     use_cache=True,
+...     num_beams=1,
+...     bad_words_ids=[[processor.tokenizer.unk_token_id]],
+...     return_dict_in_generate=True,
+... )
+
+>>> sequence = processor.batch_decode(outputs.sequences)[0]
+>>> sequence = sequence.replace(processor.tokenizer.eos_token, "").replace(processor.tokenizer.pad_token, "")
+>>> sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()  # remove first task start token
+>>> print(processor.token2json(sequence))
+{'menu': {'nm': 'CINNAMON SUGAR', 'unitprice': '17,000', 'cnt': '1 x', 'price': '17,000'}, 'sub_total': {'subtotal_price': '17,000'}, 'total': {'total_price': '17,000', 'cashprice': '20,000', 'changeprice': '3,000'}}
+```
+
+- Step-by-step Document Visual Question Answering (DocVQA)
+
+```py
+>>> import re
+
+>>> from transformers import DonutProcessor, VisionEncoderDecoderModel
+>>> from datasets import load_dataset
+>>> import torch
+
+>>> processor = DonutProcessor.from_pretrained("naver-clova-ix/donut-base-finetuned-docvqa")
+>>> model = VisionEncoderDecoderModel.from_pretrained("naver-clova-ix/donut-base-finetuned-docvqa")
+
+>>> device = "cuda" if torch.cuda.is_available() else "cpu"
+>>> model.to(device)  # doctest: +IGNORE_RESULT
+
+>>> # load document image from the DocVQA dataset
+>>> dataset = load_dataset("hf-internal-testing/example-documents", split="test")
+>>> image = dataset[0]["image"]
+
+>>> # prepare decoder inputs
+>>> task_prompt = "<s_docvqa><s_question>{user_input}</s_question><s_answer>"
+>>> question = "When is the coffee break?"
+>>> prompt = task_prompt.replace("{user_input}", question)
+>>> decoder_input_ids = processor.tokenizer(prompt, add_special_tokens=False, return_tensors="pt").input_ids
+
+>>> pixel_values = processor(image, return_tensors="pt").pixel_values
+
+>>> outputs = model.generate(
+...     pixel_values.to(device),
+...     decoder_input_ids=decoder_input_ids.to(device),
+...     max_length=model.decoder.config.max_position_embeddings,
+...     early_stopping=True,
+...     pad_token_id=processor.tokenizer.pad_token_id,
+...     eos_token_id=processor.tokenizer.eos_token_id,
+...     use_cache=True,
+...     num_beams=1,
+...     bad_words_ids=[[processor.tokenizer.unk_token_id]],
+...     return_dict_in_generate=True,
+... )
+
+>>> sequence = processor.batch_decode(outputs.sequences)[0]
+>>> sequence = sequence.replace(processor.tokenizer.eos_token, "").replace(processor.tokenizer.pad_token, "")
+>>> sequence = re.sub(r"<.*?>", "", sequence, count=1).strip()  # remove first task start token
+>>> print(processor.token2json(sequence))
+{'question': 'When is the coffee break?', 'answer': '11-14 to 11:39 a.m.'}
+```
+
+See the [model hub](https://huggingface.co/models?filter=donut) to look for Donut checkpoints.
+
+## Training
+
+We refer to the [tutorial notebooks](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/Donut).
+
+## DonutSwinConfig
+
+[[autodoc]] DonutSwinConfig
+
+## DonutFeatureExtractor
+
+[[autodoc]] DonutFeatureExtractor
+    - __call__
+
+## DonutProcessor
+
+[[autodoc]] DonutProcessor
+    - __call__
+    - from_pretrained
+    - save_pretrained
+    - batch_decode
+    - decode
+
+## DonutSwinModel
+
+[[autodoc]] DonutSwinModel
+    - forward

docs/source/en/model_doc/encoder-decoder.mdx

@@ -12,6 +12,8 @@ specific language governing permissions and limitations under the License.
 
 # Encoder Decoder Models
 
+## Overview
+
 The [`EncoderDecoderModel`] can be used to initialize a sequence-to-sequence model with any
 pretrained autoencoding model as the encoder and any pretrained autoregressive model as the decoder.
 
@@ -25,15 +27,77 @@ any other models (see the examples for more information).
 An application of this architecture could be to leverage two pretrained [`BertModel`] as the encoder
 and decoder for a summarization model as was shown in: [Text Summarization with Pretrained Encoders](https://arxiv.org/abs/1908.08345) by Yang Liu and Mirella Lapata.
 
-The [`~TFEncoderDecoderModel.from_pretrained`] currently doesn't support initializing the model from a
+## Randomly initializing `EncoderDecoderModel` from model configurations.
+
+[`EncoderDecoderModel`] can be randomly initialized from an encoder and a decoder config. In the following example, we show how to do this using the default [`BertModel`] configuration for the encoder and the default [`BertForCausalLM`] configuration for the decoder.
+
+```python
+>>> from transformers import BertConfig, EncoderDecoderConfig, EncoderDecoderModel
+
+>>> config_encoder = BertConfig()
+>>> config_decoder = BertConfig()
+
+>>> config = EncoderDecoderConfig.from_encoder_decoder_configs(config_encoder, config_decoder)
+>>> model = EncoderDecoderModel(config=config)
+```
+
+## Initialising `EncoderDecoderModel` from a pretrained encoder and a pretrained decoder.
+
+[`EncoderDecoderModel`] can be initialized from a pretrained encoder checkpoint and a pretrained decoder checkpoint. Note that any pretrained auto-encoding model, *e.g.* BERT, can serve as the encoder and both pretrained auto-encoding models, *e.g.* BERT, pretrained causal language models, *e.g.* GPT2, as well as the pretrained decoder part of sequence-to-sequence models, *e.g.* decoder of BART, can be used as the decoder.
+Depending on which architecture you choose as the decoder, the cross-attention layers might be randomly initialized.
+Initializing [`EncoderDecoderModel`] from a pretrained encoder and decoder checkpoint requires the model to be fine-tuned on a downstream task, as has been shown in [the *Warm-starting-encoder-decoder blog post*](https://huggingface.co/blog/warm-starting-encoder-decoder).
+To do so, the `EncoderDecoderModel` class provides a [`EncoderDecoderModel.from_encoder_decoder_pretrained`] method.
+
+```python
+>>> from transformers import EncoderDecoderModel, BertTokenizer
+
+>>> tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+>>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("bert-base-uncased", "bert-base-uncased")
+```
+
+## Loading an existing `EncoderDecoderModel` checkpoint and perform inference.
+
+To load fine-tuned checkpoints of the `EncoderDecoderModel` class, [`EncoderDecoderModel`] provides the `from_pretrained(...)` method just like any other model architecture in Transformers.
+
+To perform inference, one uses the [`generate`] method, which allows to autoregressively generate text. This method supports various forms of decoding, such as greedy, beam search and multinomial sampling.
+
+```python
+>>> from transformers import AutoTokenizer, EncoderDecoderModel
+
+>>> # load a fine-tuned seq2seq model and corresponding tokenizer
+>>> model = EncoderDecoderModel.from_pretrained("patrickvonplaten/bert2bert_cnn_daily_mail")
+>>> tokenizer = AutoTokenizer.from_pretrained("patrickvonplaten/bert2bert_cnn_daily_mail")
+
+>>> # let's perform inference on a long piece of text
+>>> ARTICLE_TO_SUMMARIZE = (
+...     "PG&E stated it scheduled the blackouts in response to forecasts for high winds "
+...     "amid dry conditions. The aim is to reduce the risk of wildfires. Nearly 800 thousand customers were "
+...     "scheduled to be affected by the shutoffs which were expected to last through at least midday tomorrow."
+... )
+>>> input_ids = tokenizer(ARTICLE_TO_SUMMARIZE, return_tensors="pt").input_ids
+
+>>> # autoregressively generate summary (uses greedy decoding by default)
+>>> generated_ids = model.generate(input_ids)
+>>> generated_text = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+>>> print(generated_text)
+nearly 800 thousand customers were affected by the shutoffs. the aim is to reduce the risk of wildfires. nearly 800, 000 customers were expected to be affected by high winds amid dry conditions. pg & e said it scheduled the blackouts to last through at least midday tomorrow.
+```
+
+## Loading a PyTorch checkpoint into `TFEncoderDecoderModel`.
+
+[`TFEncoderDecoderModel.from_pretrained`] currently doesn't support initializing the model from a
 pytorch checkpoint. Passing `from_pt=True` to this method will throw an exception. If there are only pytorch
 checkpoints for a particular encoder-decoder model, a workaround is:
 
 ```python
 >>> # a workaround to load from pytorch checkpoint
+>>> from transformers import EncoderDecoderModel, TFEncoderDecoderModel
+
 >>> _model = EncoderDecoderModel.from_pretrained("patrickvonplaten/bert2bert-cnn_dailymail-fp16")
+
 >>> _model.encoder.save_pretrained("./encoder")
 >>> _model.decoder.save_pretrained("./decoder")
+
 >>> model = TFEncoderDecoderModel.from_encoder_decoder_pretrained(
 ...     "./encoder", "./decoder", encoder_from_pt=True, decoder_from_pt=True
 ... )
@@ -41,6 +105,38 @@ checkpoints for a particular encoder-decoder model, a workaround is:
 >>> model.config = _model.config
 ```
 
+## Training
+
+Once the model is created, it can be fine-tuned similar to BART, T5 or any other encoder-decoder model.
+As you can see, only 2 inputs are required for the model in order to compute a loss: `input_ids` (which are the
+`input_ids` of the encoded input sequence) and `labels` (which are the `input_ids` of the encoded
+target sequence).
+
+```python
+>>> from transformers import BertTokenizer, EncoderDecoderModel
+
+>>> tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+>>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("bert-base-uncased", "bert-base-uncased")
+
+>>> model.config.decoder_start_token_id = tokenizer.cls_token_id
+>>> model.config.pad_token_id = tokenizer.pad_token_id
+
+>>> input_ids = tokenizer(
+...     "The tower is 324 metres (1,063 ft) tall, about the same height as an 81-storey building, and the tallest structure in Paris. Its base is square, measuring 125 metres (410 ft) on each side.During its construction, the Eiffel Tower surpassed the Washington Monument to become the tallest man-made structure in the world, a title it held for 41 years until the Chrysler Building in New York City was  finished in 1930. It was the first structure to reach a height of 300 metres. Due to the addition of a broadcasting aerial at the top of the tower in 1957, it is now taller than the Chrysler Building by 5.2 metres (17 ft).Excluding transmitters, the Eiffel Tower is the second tallest free-standing structure in France after the Millau Viaduct.",
+...     return_tensors="pt",
+... ).input_ids
+
+>>> labels = tokenizer(
+...     "the eiffel tower surpassed the washington monument to become the tallest structure in the world. it was the first structure to reach a height of 300 metres in paris in 1930. it is now taller than the chrysler building by 5. 2 metres ( 17 ft ) and is the second tallest free - standing structure in paris.",
+...     return_tensors="pt",
+... ).input_ids
+
+>>> # the forward function automatically creates the correct decoder_input_ids
+>>> loss = model(input_ids=input_ids, labels=labels).loss
+```
+
+Detailed [colab](https://colab.research.google.com/drive/1WIk2bxglElfZewOHboPFNj8H44_VAyKE?usp=sharing#scrollTo=ZwQIEhKOrJpl) for training.
+
 This model was contributed by [thomwolf](https://github.com/thomwolf). This model's TensorFlow and Flax versions
 were contributed by [ydshieh](https://github.com/ydshieh).
 

docs/source/en/model_doc/ernie.mdx

@@ -0,0 +1,102 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# ERNIE
+
+## Overview
+ERNIE is a series of powerful models proposed by baidu, especially in Chinese tasks,
+including [ERNIE1.0](https://arxiv.org/abs/1904.09223), [ERNIE2.0](https://ojs.aaai.org/index.php/AAAI/article/view/6428),
+[ERNIE3.0](https://arxiv.org/abs/2107.02137), [ERNIE-Gram](https://arxiv.org/abs/2010.12148), [ERNIE-health](https://arxiv.org/abs/2110.07244), etc.
+
+These models are contributed by [nghuyong](https://huggingface.co/nghuyong) and the official code can be found in [PaddleNLP](https://github.com/PaddlePaddle/PaddleNLP) (in PaddlePaddle).
+
+### How to use
+Take `ernie-1.0-base-zh` as an example:
+
+```Python
+from transformers import AutoTokenizer, AutoModel
+tokenizer = AutoTokenizer.from_pretrained("nghuyong/ernie-1.0-base-zh")
+model = AutoModel.from_pretrained("nghuyong/ernie-1.0-base-zh")
+```
+
+### Supported Models
+
+|     Model Name      | Language |           Description           |
+|:-------------------:|:--------:|:-------------------------------:|
+|  ernie-1.0-base-zh  | Chinese  | Layer:12, Heads:12, Hidden:768  |
+|  ernie-2.0-base-en  | English  | Layer:12, Heads:12, Hidden:768  |
+| ernie-2.0-large-en  | English  | Layer:24, Heads:16, Hidden:1024 |
+|  ernie-3.0-base-zh  | Chinese  | Layer:12, Heads:12, Hidden:768  |
+| ernie-3.0-medium-zh | Chinese  |  Layer:6, Heads:12, Hidden:768  |
+|  ernie-3.0-mini-zh  | Chinese  |  Layer:6, Heads:12, Hidden:384  |
+| ernie-3.0-micro-zh  | Chinese  |  Layer:4, Heads:12, Hidden:384  |
+|  ernie-3.0-nano-zh  | Chinese  |  Layer:4, Heads:12, Hidden:312  |
+|   ernie-health-zh   | Chinese  | Layer:12, Heads:12, Hidden:768  |
+|    ernie-gram-zh    | Chinese  | Layer:12, Heads:12, Hidden:768  |
+
+You can find all the supported models from huggingface's model hub: [huggingface.co/nghuyong](https://huggingface.co/nghuyong), and model details from paddle's official
+repo: [PaddleNLP](https://paddlenlp.readthedocs.io/zh/latest/model_zoo/transformers/ERNIE/contents.html)
+and [ERNIE](https://github.com/PaddlePaddle/ERNIE/blob/repro).
+
+## ErnieConfig
+
+[[autodoc]] ErnieConfig
+    - all
+
+## Ernie specific outputs
+
+[[autodoc]] models.ernie.modeling_ernie.ErnieForPreTrainingOutput
+
+## ErnieModel
+
+[[autodoc]] ErnieModel
+    - forward
+
+## ErnieForPreTraining
+
+[[autodoc]] ErnieForPreTraining
+    - forward
+
+## ErnieForCausalLM
+
+[[autodoc]] ErnieForCausalLM
+    - forward
+
+## ErnieForMaskedLM
+
+[[autodoc]] ErnieForMaskedLM
+    - forward
+
+## ErnieForNextSentencePrediction
+
+[[autodoc]] ErnieForNextSentencePrediction
+    - forward
+
+## ErnieForSequenceClassification
+
+[[autodoc]] ErnieForSequenceClassification
+    - forward
+
+## ErnieForMultipleChoice
+
+[[autodoc]] ErnieForMultipleChoice
+    - forward
+
+## ErnieForTokenClassification
+
+[[autodoc]] ErnieForTokenClassification
+    - forward
+
+## ErnieForQuestionAnswering
+
+[[autodoc]] ErnieForQuestionAnswering
+    - forward

docs/source/en/model_doc/gpt_neox.mdx

@@ -0,0 +1,76 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# GPT-NeoX
+
+## Overview
+
+We introduce GPT-NeoX-20B, a 20 billion parameter autoregressive language model trained on the Pile, whose weights will
+be made freely and openly available to the public through a permissive license. It is, to the best of our knowledge,
+the largest dense autoregressive model that has publicly available weights at the time of submission. In this work,
+we describe GPT-NeoX-20B's architecture and training and evaluate its performance on a range of language-understanding,
+mathematics, and knowledge-based tasks. We find that GPT-NeoX-20B is a particularly powerful few-shot reasoner and
+gains far more in performance when evaluated five-shot than similarly sized GPT-3 and FairSeq models. We open-source
+the training and evaluation code, as well as the model weights, at [https://github.com/EleutherAI/gpt-neox](https://github.com/EleutherAI/gpt-neox).
+
+Development of the model was led by Sid Black, Stella Biderman and Eric Hallahan, and the model was trained with
+generous the support of [CoreWeave](https://www.coreweave.com/).
+
+GPT-NeoX-20B was trained with fp16, thus it is recommended to initialize the model as follows:
+
+```python
+model = GPTNeoXForCausalLM.from_pretrained("EleutherAI/gpt-neox-20b").half().cuda()
+```
+
+GPT-NeoX-20B also has a different tokenizer from the one used in GPT-J-6B and GPT-Neo. The new tokenizer allocates
+additional tokens to whitespace characters, making the model more suitable for certain tasks like code generation.
+
+### Generation
+
+The `generate()` method can be used to generate text using GPT Neo model.
+
+```python
+>>> from transformers import GPTNeoXForCausalLM, GPTNeoXTokenizerFast
+
+>>> model = GPTNeoXForCausalLM.from_pretrained("EleutherAI/gpt-neox-20b")
+>>> tokenizer = GPTNeoXTokenizerFast.from_pretrained("EleutherAI/gpt-neox-20b")
+
+>>> prompt = "GPTNeoX20B is a 20B-parameter autoregressive Transformer model developed by EleutherAI."
+
+>>> input_ids = tokenizer(prompt, return_tensors="pt").input_ids
+
+>>> gen_tokens = model.generate(
+...     input_ids,
+...     do_sample=True,
+...     temperature=0.9,
+...     max_length=100,
+... )
+>>> gen_text = tokenizer.batch_decode(gen_tokens)[0]
+```
+
+## GPTNeoXConfig
+
+[[autodoc]] GPTNeoXConfig
+
+## GPTNeoXTokenizerFast
+
+[[autodoc]] GPTNeoXTokenizerFast
+
+## GPTNeoXModel
+
+[[autodoc]] GPTNeoXModel
+    - forward
+
+## GPTNeoXForCausalLM
+
+[[autodoc]] GPTNeoXForCausalLM
+    - forward

docs/source/en/model_doc/groupvit.mdx

@@ -0,0 +1,61 @@
+<!--Copyright 2022 NVIDIA and The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# GroupViT
+
+## Overview
+
+The GroupViT model was proposed in [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
+Inspired by [CLIP](clip), GroupViT is a vision-language model that can perform zero-shot semantic segmentation on any given vocabulary categories.
+
+The abstract from the paper is the following:
+
+*Grouping and recognition are important components of visual scene understanding, e.g., for object detection and semantic segmentation. With end-to-end deep learning systems, grouping of image regions usually happens implicitly via top-down supervision from pixel-level recognition labels. Instead, in this paper, we propose to bring back the grouping mechanism into deep networks, which allows semantic segments to emerge automatically with only text supervision. We propose a hierarchical Grouping Vision Transformer (GroupViT), which goes beyond the regular grid structure representation and learns to group image regions into progressively larger arbitrary-shaped segments. We train GroupViT jointly with a text encoder on a large-scale image-text dataset via contrastive losses. With only text supervision and without any pixel-level annotations, GroupViT learns to group together semantic regions and successfully transfers to the task of semantic segmentation in a zero-shot manner, i.e., without any further fine-tuning. It achieves a zero-shot accuracy of 52.3% mIoU on the PASCAL VOC 2012 and 22.4% mIoU on PASCAL Context datasets, and performs competitively to state-of-the-art transfer-learning methods requiring greater levels of supervision.*
+
+Tips:
+
+- You may specify `output_segmentation=True` in the forward of `GroupViTModel` to get the segmentation logits of input texts. 
+- The quickest way to get started with GroupViT is by checking the [example notebooks](https://github.com/xvjiarui/GroupViT/blob/main/demo/GroupViT_hf_inference_notebook.ipynb) (which showcase zero-shot segmentation inference). One can also check out the [HuggingFace Spaces demo](https://huggingface.co/spaces/xvjiarui/GroupViT) to play with GroupViT. 
+
+This model was contributed by [xvjiarui](https://huggingface.co/xvjiarui).
+The original code can be found [here](https://github.com/NVlabs/GroupViT).
+
+
+## GroupViTConfig
+
+[[autodoc]] GroupViTConfig
+    - from_text_vision_configs
+
+## GroupViTTextConfig
+
+[[autodoc]] GroupViTTextConfig
+
+## GroupViTVisionConfig
+
+[[autodoc]] GroupViTVisionConfig
+
+## GroupViTModel
+
+[[autodoc]] GroupViTModel
+    - forward
+    - get_text_features
+    - get_image_features
+
+## GroupViTTextModel
+
+[[autodoc]] GroupViTTextModel
+    - forward
+
+## GroupViTVisionModel
+
+[[autodoc]] GroupViTVisionModel
+    - forward

docs/source/en/model_doc/layoutlm.mdx

@@ -67,7 +67,8 @@ occurs. Those can be obtained using the Python Image Library (PIL) library for e
 ```python
 from PIL import Image
 
-image = Image.open("name_of_your_document - can be a png file, pdf, etc.")
+# Document can be a png, jpg, etc. PDFs must be converted to images.
+image = Image.open(name_of_your_document).convert("RGB")
 
 width, height = image.size
 ```
@@ -107,6 +108,10 @@ This model was contributed by [liminghao1630](https://huggingface.co/liminghao16
 
 [[autodoc]] LayoutLMForTokenClassification
 
+## LayoutLMForQuestionAnswering
+
+[[autodoc]] LayoutLMForQuestionAnswering
+
 ## TFLayoutLMModel
 
 [[autodoc]] TFLayoutLMModel
@@ -122,3 +127,7 @@ This model was contributed by [liminghao1630](https://huggingface.co/liminghao16
 ## TFLayoutLMForTokenClassification
 
 [[autodoc]] TFLayoutLMForTokenClassification
+
+## TFLayoutLMForQuestionAnswering
+
+[[autodoc]] TFLayoutLMForQuestionAnswering

docs/source/en/model_doc/layoutlmv2.mdx

@@ -44,6 +44,14 @@ including FUNSD (0.7895 -> 0.8420), CORD (0.9493 -> 0.9601), SROIE (0.9524 -> 0.
 RVL-CDIP (0.9443 -> 0.9564), and DocVQA (0.7295 -> 0.8672). The pre-trained LayoutLMv2 model is publicly available at
 this https URL.*
 
+LayoutLMv2 depends on `detectron2`, `torchvision` and `tesseract`. Run the
+following to install them: 
+```
+python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
+python -m pip install torchvision tesseract
+```
+(If you are developing for LayoutLMv2, note that passing the doctests also requires the installation of these packages.)
+
 Tips:
 
 - The main difference between LayoutLMv1 and LayoutLMv2 is that the latter incorporates visual embeddings during

docs/source/en/model_doc/layoutlmv3.mdx

@@ -0,0 +1,106 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# LayoutLMv3
+
+## Overview
+
+The LayoutLMv3 model was proposed in [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
+LayoutLMv3 simplifies [LayoutLMv2](layoutlmv2) by using patch embeddings (as in [ViT](vit)) instead of leveraging a CNN backbone, and pre-trains the model on 3 objectives: masked language modeling (MLM), masked image modeling (MIM)
+and word-patch alignment (WPA).
+
+The abstract from the paper is the following:
+
+*Self-supervised pre-training techniques have achieved remarkable progress in Document AI. Most multimodal pre-trained models use a masked language modeling objective to learn bidirectional representations on the text modality, but they differ in pre-training objectives for the image modality. This discrepancy adds difficulty to multimodal representation learning. In this paper, we propose LayoutLMv3 to pre-train multimodal Transformers for Document AI with unified text and image masking. Additionally, LayoutLMv3 is pre-trained with a word-patch alignment objective to learn cross-modal alignment by predicting whether the corresponding image patch of a text word is masked. The simple unified architecture and training objectives make LayoutLMv3 a general-purpose pre-trained model for both text-centric and image-centric Document AI tasks. Experimental results show that LayoutLMv3 achieves state-of-the-art performance not only in text-centric tasks, including form understanding, receipt understanding, and document visual question answering, but also in image-centric tasks such as document image classification and document layout analysis.*
+
+Tips:
+
+- In terms of data processing, LayoutLMv3 is identical to its predecessor [LayoutLMv2](layoutlmv2), except that:
+    - images need to be resized and normalized with channels in regular RGB format. LayoutLMv2 on the other hand normalizes the images internally and expects the channels in BGR format.
+    - text is tokenized using byte-pair encoding (BPE), as opposed to WordPiece.
+  Due to these differences in data preprocessing, one can use [`LayoutLMv3Processor`] which internally combines a [`LayoutLMv3FeatureExtractor`] (for the image modality) and a [`LayoutLMv3Tokenizer`]/[`LayoutLMv3TokenizerFast`] (for the text modality) to prepare all data for the model.
+- Regarding usage of [`LayoutLMv3Processor`], we refer to the [usage guide](layoutlmv2#usage-layoutlmv2processor) of its predecessor.
+- Demo notebooks for LayoutLMv3 can be found [here](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/LayoutLMv3).
+- Demo scripts can be found [here](https://github.com/huggingface/transformers/tree/main/examples/research_projects/layoutlmv3).
+
+<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/layoutlmv3_architecture.png"
+alt="drawing" width="600"/>
+
+<small> LayoutLMv3 architecture. Taken from the <a href="https://arxiv.org/abs/2204.08387">original paper</a>. </small>
+
+This model was contributed by [nielsr](https://huggingface.co/nielsr). The TensorFlow version of this model was added by [chriskoo](https://huggingface.co/chriskoo), [tokec](https://huggingface.co/tokec), and [lre](https://huggingface.co/lre). The original code can be found [here](https://github.com/microsoft/unilm/tree/master/layoutlmv3).
+
+
+## LayoutLMv3Config
+
+[[autodoc]] LayoutLMv3Config
+
+## LayoutLMv3FeatureExtractor
+
+[[autodoc]] LayoutLMv3FeatureExtractor
+    - __call__
+
+## LayoutLMv3Tokenizer
+
+[[autodoc]] LayoutLMv3Tokenizer
+    - __call__
+    - save_vocabulary
+
+## LayoutLMv3TokenizerFast
+
+[[autodoc]] LayoutLMv3TokenizerFast
+    - __call__
+
+## LayoutLMv3Processor
+
+[[autodoc]] LayoutLMv3Processor
+    - __call__
+
+## LayoutLMv3Model
+
+[[autodoc]] LayoutLMv3Model
+    - forward
+
+## LayoutLMv3ForSequenceClassification
+
+[[autodoc]] LayoutLMv3ForSequenceClassification
+    - forward
+
+## LayoutLMv3ForTokenClassification
+
+[[autodoc]] LayoutLMv3ForTokenClassification
+    - forward
+
+## LayoutLMv3ForQuestionAnswering
+
+[[autodoc]] LayoutLMv3ForQuestionAnswering
+    - forward
+
+## TFLayoutLMv3Model
+
+[[autodoc]] TFLayoutLMv3Model
+    - call
+
+## TFLayoutLMv3ForSequenceClassification
+
+[[autodoc]] TFLayoutLMv3ForSequenceClassification
+    - call
+
+## TFLayoutLMv3ForTokenClassification
+
+[[autodoc]] TFLayoutLMv3ForTokenClassification
+    - call
+
+## TFLayoutLMv3ForQuestionAnswering
+
+[[autodoc]] TFLayoutLMv3ForQuestionAnswering
+    - call

docs/source/en/model_doc/led.mdx

@@ -44,8 +44,10 @@ Tips:
 - LED makes use of *global attention* by means of the `global_attention_mask` (see
   [`LongformerModel`]). For summarization, it is advised to put *global attention* only on the first
   `<s>` token. For question answering, it is advised to put *global attention* on all tokens of the question.
-- To fine-tune LED on all 16384, it is necessary to enable *gradient checkpointing* by executing
-  `model.gradient_checkpointing_enable()`.
+- To fine-tune LED on all 16384, *gradient checkpointing* can be enabled in case training leads to out-of-memory (OOM)
+  errors. This can be done by executing `model.gradient_checkpointing_enable()`. 
+ Moreover, the `use_cache=False`
+  flag can be used to disable the caching mechanism to save memory.
 - A notebook showing how to evaluate LED, can be accessed [here](https://colab.research.google.com/drive/12INTTR6n64TzS4RrXZxMSXfrOd9Xzamo?usp=sharing).
 - A notebook showing how to fine-tune LED, can be accessed [here](https://colab.research.google.com/drive/12LjJazBl7Gam0XBPy_y0CTOJZeZ34c2v?usp=sharing).
 

docs/source/en/model_doc/levit.mdx

@@ -0,0 +1,87 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# LeViT
+
+## Overview
+
+The LeViT model was proposed in [LeViT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervé Jégou, Matthijs Douze. LeViT improves the [Vision Transformer (ViT)](vit) in performance and efficiency by a few architectural differences such as activation maps with decreasing resolutions in Transformers and the introduction of an attention bias to integrate positional information.
+
+The abstract from the paper is the following:
+
+*We design a family of image classification architectures that optimize the trade-off between accuracy
+and efficiency in a high-speed regime. Our work exploits recent findings in attention-based architectures, 
+which are competitive on highly parallel processing hardware. We revisit principles from the extensive 
+literature on convolutional neural networks to apply them to transformers, in particular activation maps 
+with decreasing resolutions. We also introduce the attention bias, a new way to integrate positional information
+in vision transformers. As a result, we propose LeVIT: a hybrid neural network for fast inference image classification. 
+We consider different measures of efficiency on different hardware platforms, so as to best reflect a wide range of 
+application scenarios. Our extensive experiments empirically validate our technical choices and show they are suitable 
+to most architectures. Overall, LeViT significantly outperforms existing convnets and vision transformers with respect 
+to the speed/accuracy tradeoff. For example, at 80% ImageNet top-1 accuracy, LeViT is 5 times faster than EfficientNet on CPU. *
+
+<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/levit_architecture.png"
+alt="drawing" width="600"/> 
+
+<small> LeViT Architecture. Taken from the <a href="https://arxiv.org/abs/2104.01136">original paper</a>.</small>
+
+Tips:
+
+- Compared to ViT, LeViT models use an additional distillation head to effectively learn from a teacher (which, in the LeViT paper, is a ResNet like-model). The distillation head is learned through backpropagation under supervision of a ResNet like-model. They also draw inspiration from convolution neural networks to use activation maps with decreasing resolutions to increase the efficiency.
+- There are 2 ways to fine-tune distilled models, either (1) in a classic way, by only placing a prediction head on top
+  of the final hidden state and not using the distillation head, or (2) by placing both a prediction head and distillation 
+  head on top of the final hidden state. In that case, the prediction head is trained using regular cross-entropy between 
+  the prediction of the head and the ground-truth label, while the distillation prediction head is trained using hard distillation 
+  (cross-entropy between the prediction of the distillation head and the label predicted by the teacher). At inference time, 
+  one takes the average prediction between both heads as final prediction. (2) is also called "fine-tuning with distillation", 
+  because one relies on a teacher that has already been fine-tuned on the downstream dataset. In terms of models, (1) corresponds 
+  to [`LevitForImageClassification`] and (2) corresponds to [`LevitForImageClassificationWithTeacher`].
+- All released checkpoints were pre-trained and fine-tuned on  [ImageNet-1k](https://huggingface.co/datasets/imagenet-1k) 
+  (also referred to as ILSVRC 2012, a collection of 1.3 million images and 1,000 classes). only. No external data was used. This is in
+  contrast with the original ViT model, which used external data like the JFT-300M dataset/Imagenet-21k for
+  pre-training.
+- The authors of LeViT released 5 trained LeViT models, which you can directly plug into [`LevitModel`] or [`LevitForImageClassification`]. 
+  Techniques like data augmentation, optimization, and regularization were used in order to simulate training on a much larger dataset
+  (while only using ImageNet-1k for pre-training). The 5 variants available are (all trained on images of size 224x224):
+  *facebook/levit-128S*, *facebook/levit-128*, *facebook/levit-192*, *facebook/levit-256* and
+  *facebook/levit-384*. Note that one should use [`LevitFeatureExtractor`] in order to
+  prepare images for the model.
+- [`LevitForImageClassificationWithTeacher`] currently supports only inference and not training or fine-tuning.
+- You can check out demo notebooks regarding inference as well as fine-tuning on custom data [here](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/VisionTransformer) 
+  (you can just replace [`ViTFeatureExtractor`] by [`LevitFeatureExtractor`] and [`ViTForImageClassification`] by [`LevitForImageClassification`] or [`LevitForImageClassificationWithTeacher`]).
+
+This model was contributed by [anugunj](https://huggingface.co/anugunj). The original code can be found [here](https://github.com/facebookresearch/LeViT).
+
+
+## LevitConfig
+
+[[autodoc]] LevitConfig
+
+## LevitFeatureExtractor
+
+[[autodoc]] LevitFeatureExtractor
+    - __call__
+
+## LevitModel
+
+[[autodoc]] LevitModel
+    - forward
+
+## LevitForImageClassification
+
+[[autodoc]] LevitForImageClassification
+    - forward
+
+## LevitForImageClassificationWithTeacher
+
+[[autodoc]] LevitForImageClassificationWithTeacher
+    - forward

docs/source/en/model_doc/longt5.mdx

@@ -0,0 +1,121 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# LongT5
+
+## Overview
+
+The LongT5 model was proposed in [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916)
+by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung and Yinfei Yang. It's an
+encoder-decoder transformer pre-trained in a text-to-text denoising generative setting. LongT5 model is an extension of
+T5 model, and it enables using one of the two different efficient attention mechanisms - (1) Local attention, or (2)
+Transient-Global attention.
+
+
+The abstract from the paper is the following:
+
+*Recent work has shown that either (1) increasing the input length or (2) increasing model size can improve the
+performance of Transformer-based neural models. In this paper, we present a new model, called LongT5, with which we
+explore the effects of scaling both the input length and model size at the same time. Specifically, we integrated
+attention ideas from long-input transformers (ETC), and adopted pre-training strategies from summarization pre-training
+(PEGASUS) into the scalable T5 architecture. The result is a new attention mechanism we call {\em Transient Global}
+(TGlobal), which mimics ETC's local/global attention mechanism, but without requiring additional side-inputs. We are
+able to achieve state-of-the-art results on several summarization tasks and outperform the original T5 models on
+question answering tasks.*
+
+Tips:
+
+- [`LongT5ForConditionalGeneration`] is an extension of [`T5ForConditionalGeneration`] exchanging the traditional
+encoder *self-attention* layer with efficient either *local* attention or *transient-global* (*tglobal*) attention.
+- Unlike the T5 model, LongT5 does not use a task prefix. Furthermore, it uses a different pre-training objective
+inspired by the pre-training of [`PegasusForConditionalGeneration`].
+- LongT5 model is designed to work efficiently and very well on long-range *sequence-to-sequence* tasks where the
+input sequence exceeds commonly used 512 tokens. It is capable of handling input sequences of a length up to 16,384 tokens.
+- For *Local Attention*, the sparse sliding-window local attention operation allows a given token to attend only `r`
+tokens to the left and right of it (with `r=127` by default). *Local Attention* does not introduce any new parameters
+to the model. The complexity of the mechanism is linear in input sequence length `l`: `O(l*r)`.
+- *Transient Global Attention* is an extension of the *Local Attention*. It, furthermore, allows each input token to
+interact with all other tokens in the layer. This is achieved via splitting an input sequence into blocks of a fixed
+length `k` (with a default `k=16`). Then, a global token for such a block is obtained via summing and normalizing the embeddings of every token
+in the block. Thanks to this, the attention allows each token to attend to both nearby tokens like in Local attention, and
+also every global token like in the case of standard global attention (*transient* represents the fact the global tokens
+are constructed dynamically within each attention operation).  As a consequence, *TGlobal* attention introduces
+a few new parameters -- global relative position biases and a layer normalization for global token's embedding.
+The complexity of this mechanism is `O(l(r + l/k))`.
+- An example showing how to evaluate a fine-tuned LongT5 model on the [pubmed dataset](https://huggingface.co/datasets/scientific_papers) is below.
+
+```python
+>>> import evaluate
+>>> from datasets import load_dataset
+>>> from transformers import AutoTokenizer, LongT5ForConditionalGeneration
+
+>>> dataset = load_dataset("scientific_papers", "pubmed", split="validation")
+>>> model = (
+...     LongT5ForConditionalGeneration.from_pretrained("Stancld/longt5-tglobal-large-16384-pubmed-3k_steps")
+...     .to("cuda")
+...     .half()
+... )
+>>> tokenizer = AutoTokenizer.from_pretrained("Stancld/longt5-tglobal-large-16384-pubmed-3k_steps")
+
+
+>>> def generate_answers(batch):
+...     inputs_dict = tokenizer(
+...         batch["article"], max_length=16384, padding="max_length", truncation=True, return_tensors="pt"
+...     )
+...     input_ids = inputs_dict.input_ids.to("cuda")
+...     attention_mask = inputs_dict.attention_mask.to("cuda")
+...     output_ids = model.generate(input_ids, attention_mask=attention_mask, max_length=512, num_beams=2)
+...     batch["predicted_abstract"] = tokenizer.batch_decode(output_ids, skip_special_tokens=True)
+...     return batch
+
+
+>>> result = dataset.map(generate_answer, batched=True, batch_size=2)
+>>> rouge = evaluate.load("rouge")
+>>> rouge.compute(predictions=result["predicted_abstract"], references=result["abstract"])
+```
+
+This model was contributed by [stancld](https://huggingface.co/stancld).
+The original code can be found [here](https://github.com/google-research/longt5).
+
+
+## LongT5Config
+
+[[autodoc]] LongT5Config
+
+## LongT5Model
+
+[[autodoc]] LongT5Model
+    - forward
+
+## LongT5ForConditionalGeneration
+
+[[autodoc]] LongT5ForConditionalGeneration
+    - forward
+
+## LongT5EncoderModel
+
+[[autodoc]] LongT5EncoderModel
+    - forward
+
+## FlaxLongT5Model
+
+[[autodoc]] FlaxLongT5Model
+    - __call__
+    - encode
+    - decode
+
+## FlaxLongT5ForConditionalGeneration
+
+[[autodoc]] FlaxLongT5ForConditionalGeneration
+    - __call__
+    - encode
+    - decode

docs/source/en/model_doc/luke.mdx

@@ -152,3 +152,23 @@ This model was contributed by [ikuyamada](https://huggingface.co/ikuyamada) and
 
 [[autodoc]] LukeForEntitySpanClassification
     - forward
+
+## LukeForSequenceClassification
+
+[[autodoc]] LukeForSequenceClassification
+    - forward
+
+## LukeForMultipleChoice
+
+[[autodoc]] LukeForMultipleChoice
+    - forward
+
+## LukeForTokenClassification
+
+[[autodoc]] LukeForTokenClassification
+    - forward
+
+## LukeForQuestionAnswering
+
+[[autodoc]] LukeForQuestionAnswering
+    - forward

docs/source/en/model_doc/m2m_100.mdx

@@ -55,9 +55,7 @@ tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100_418M", src_lang="en
 src_text = "Life is like a box of chocolates."
 tgt_text = "La vie est comme une boîte de chocolat."
 
-model_inputs = tokenizer(src_text, return_tensors="pt")
-with tokenizer.as_target_tokenizer():
-    labels = tokenizer(tgt_text, return_tensors="pt").input_ids
+model_inputs = tokenizer(src_text, text_target=tgt_text, return_tensors="pt")
 
 loss = model(**model_inputs, labels=labels)  # forward pass
 ```

docs/source/en/model_doc/marian.mdx

@@ -155,7 +155,7 @@ Example of translating english to many romance languages, using old-style 2 char
 ## MarianTokenizer
 
 [[autodoc]] MarianTokenizer
-    - as_target_tokenizer
+    - build_inputs_with_special_tokens
 
 ## MarianModel
 

docs/source/en/model_doc/maskformer.mdx

@@ -33,7 +33,7 @@ Tips:
   `get_num_masks` function inside in the `MaskFormerLoss` class of `modeling_maskformer.py`. When training on multiple nodes, this should be
   set to the average number of target masks across all nodes, as can be seen in the original implementation [here](https://github.com/facebookresearch/MaskFormer/blob/da3e60d85fdeedcb31476b5edd7d328826ce56cc/mask_former/modeling/criterion.py#L169).
 - One can use [`MaskFormerFeatureExtractor`] to prepare images for the model and optional targets for the model.
-- To get the final segmentation, depending on the task, you can call [`~MaskFormerFeatureExtractor.post_process_semantic_segmentation`] or [`~MaskFormerFeatureExtractor.post_process_panoptic_segmentation`]. Both tasks can be solved using [`MaskFormerForInstanceSegmentation`] output, the latter needs an additional `is_thing_map` to know which instances must be merged together..
+- To get the final segmentation, depending on the task, you can call [`~MaskFormerFeatureExtractor.post_process_semantic_segmentation`] or [`~MaskFormerFeatureExtractor.post_process_panoptic_segmentation`]. Both tasks can be solved using [`MaskFormerForInstanceSegmentation`] output, panoptic segmentation accepts an optional `label_ids_to_fuse` argument to fuse instances of the target object/s (e.g. sky) together.
 
 The figure below illustrates the architecture of MaskFormer. Taken from the [original paper](https://arxiv.org/abs/2107.06278).
 

docs/source/en/model_doc/mbart.mdx

@@ -34,8 +34,8 @@ model is multilingual it expects the sequences in a different format. A special
 source and target text. The source text format is `X [eos, src_lang_code]` where `X` is the source text. The
 target text format is `[tgt_lang_code] X [eos]`. `bos` is never used.
 
-The regular [`~MBartTokenizer.__call__`] will encode source text format, and it should be wrapped
-inside the context manager [`~MBartTokenizer.as_target_tokenizer`] to encode target text format.
+The regular [`~MBartTokenizer.__call__`] will encode source text format passed as first argument or with the `text`
+keyword, and target text format passed with the `text_label` keyword argument.
 
 - Supervised training
 
@@ -46,13 +46,11 @@ inside the context manager [`~MBartTokenizer.as_target_tokenizer`] to encode tar
 >>> example_english_phrase = "UN Chief Says There Is No Military Solution in Syria"
 >>> expected_translation_romanian = "Şeful ONU declară că nu există o soluţie militară în Siria"
 
->>> inputs = tokenizer(example_english_phrase, return_tensors="pt")
->>> with tokenizer.as_target_tokenizer():
-...     labels = tokenizer(expected_translation_romanian, return_tensors="pt")
+>>> inputs = tokenizer(example_english_phrase, text_target=expected_translation_romanian, return_tensors="pt")
 
 >>> model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-en-ro")
 >>> # forward pass
->>> model(**inputs, labels=batch["labels"])
+>>> model(**inputs)
 ```
 
 - Generation
@@ -108,11 +106,9 @@ tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50", src_
 src_text = " UN Chief Says There Is No Military Solution in Syria"
 tgt_text = "Şeful ONU declară că nu există o soluţie militară în Siria"
 
-model_inputs = tokenizer(src_text, return_tensors="pt")
-with tokenizer.as_target_tokenizer():
-    labels = tokenizer(tgt_text, return_tensors="pt").input_ids
+model_inputs = tokenizer(src_text, text_target=tgt_text, return_tensors="pt")
 
-model(**model_inputs, labels=labels)  # forward pass
+model(**model_inputs)  # forward pass
 ```
 
 - Generation
@@ -154,7 +150,6 @@ tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
 ## MBartTokenizer
 
 [[autodoc]] MBartTokenizer
-    - as_target_tokenizer
     - build_inputs_with_special_tokens
 
 ## MBartTokenizerFast

docs/source/en/model_doc/mctct.mdx

@@ -0,0 +1,61 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# M-CTC-T
+
+## Overview
+
+The M-CTC-T model was proposed in [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert. The model is a 1B-param transformer encoder, with a CTC head over 8065 character labels and a language identification head over 60 language ID labels. It is trained on Common Voice (version 6.1, December 2020 release) and VoxPopuli. After training on Common Voice and VoxPopuli, the model is trained on Common Voice only. The labels are unnormalized character-level transcripts (punctuation and capitalization are not removed). The model takes as input Mel filterbank features from a 16Khz audio signal.
+
+The abstract from the paper is the following:
+
+*Semi-supervised learning through pseudo-labeling has become a staple of state-of-the-art monolingual
+speech recognition systems. In this work, we extend pseudo-labeling to massively multilingual speech
+recognition with 60 languages. We propose a simple pseudo-labeling recipe that works well even
+with low-resource languages: train a supervised multilingual model, fine-tune it with semi-supervised
+learning on a target language, generate pseudo-labels for that language, and train a final model using
+pseudo-labels for all languages, either from scratch or by fine-tuning. Experiments on the labeled
+Common Voice and unlabeled VoxPopuli datasets show that our recipe can yield a model with better
+performance for many languages that also transfers well to LibriSpeech.*
+
+
+
+This model was contributed by [cwkeam](https://huggingface.co/cwkeam). The original code can be found [here](https://github.com/flashlight/wav2letter/tree/main/recipes/mling_pl).
+
+## MCTCTConfig
+
+[[autodoc]] MCTCTConfig
+
+## MCTCTFeatureExtractor
+
+[[autodoc]] MCTCTFeatureExtractor
+    - __call__
+
+## MCTCTProcessor
+
+[[autodoc]] MCTCTProcessor
+    - __call__
+    - from_pretrained
+    - save_pretrained
+    - batch_decode
+    - decode
+
+
+## MCTCTModel
+
+[[autodoc]] MCTCTModel
+    - forward
+
+## MCTCTForCTC
+
+[[autodoc]] MCTCTForCTC
+    - forward

docs/source/en/model_doc/mobilevit.mdx

@@ -0,0 +1,98 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# MobileViT
+
+## Overview
+
+The MobileViT model was proposed in [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari. MobileViT introduces a new layer that replaces local processing in convolutions with global processing using transformers. 
+
+The abstract from the paper is the following:
+
+*Light-weight convolutional neural networks (CNNs) are the de-facto for mobile vision tasks. Their spatial inductive biases allow them to learn representations with fewer parameters across different vision tasks. However, these networks are spatially local. To learn global representations, self-attention-based vision trans-formers (ViTs) have been adopted. Unlike CNNs, ViTs are heavy-weight. In this paper, we ask the following question: is it possible to combine the strengths of CNNs and ViTs to build a light-weight and low latency network for mobile vision tasks? Towards this end, we introduce MobileViT, a light-weight and general-purpose vision transformer for mobile devices. MobileViT presents a different perspective for the global processing of information with transformers, i.e., transformers as convolutions. Our results show that MobileViT significantly outperforms CNN- and ViT-based networks across different tasks and datasets. On the ImageNet-1k dataset, MobileViT achieves top-1 accuracy of 78.4% with about 6 million parameters, which is 3.2% and 6.2% more accurate than MobileNetv3 (CNN-based) and DeIT (ViT-based) for a similar number of parameters. On the MS-COCO object detection task, MobileViT is 5.7% more accurate than MobileNetv3 for a similar number of parameters.*
+
+Tips:
+
+- MobileViT is more like a CNN than a Transformer model. It does not work on sequence data but on batches of images. Unlike ViT, there are no embeddings. The backbone model outputs a feature map. You can follow [this tutorial](https://keras.io/examples/vision/mobilevit) for a lightweight introduction.
+- One can use [`MobileViTFeatureExtractor`] to prepare images for the model. Note that if you do your own preprocessing, the pretrained checkpoints expect images to be in BGR pixel order (not RGB).
+- The available image classification checkpoints are pre-trained on [ImageNet-1k](https://huggingface.co/datasets/imagenet-1k) (also referred to as ILSVRC 2012, a collection of 1.3 million images and 1,000 classes).
+- The segmentation model uses a [DeepLabV3](https://arxiv.org/abs/1706.05587) head. The available semantic segmentation checkpoints are pre-trained on [PASCAL VOC](http://host.robots.ox.ac.uk/pascal/VOC/). 
+- As the name suggests MobileViT was desgined to be performant and efficient on mobile phones. The TensorFlow versions of the MobileViT models are fully compatible with [TensorFlow Lite](https://www.tensorflow.org/lite). 
+
+  You can use the following code to convert a MobileViT checkpoint (be it image classification or semantic segmentation) to generate a 
+  TensorFlow Lite model:
+
+```py
+from transformers import TFMobileViTForImageClassification
+import tensorflow as tf
+
+
+model_ckpt = "apple/mobilevit-xx-small"
+model = TFMobileViTForImageClassification.from_pretrained(model_ckpt)
+
+converter = tf.lite.TFLiteConverter.from_keras_model(model)
+converter.optimizations = [tf.lite.Optimize.DEFAULT]
+converter.target_spec.supported_ops = [
+    tf.lite.OpsSet.TFLITE_BUILTINS,
+    tf.lite.OpsSet.SELECT_TF_OPS,
+]
+tflite_model = converter.convert()
+tflite_filename = model_ckpt.split("/")[-1] + ".tflite"
+with open(tflite_filename, "wb") as f:
+    f.write(tflite_model)
+```
+
+  The resulting model will be just **about an MB** making it a good fit for mobile applications where resources and network
+  bandwidth can be constrained. 
+
+
+This model was contributed by [matthijs](https://huggingface.co/Matthijs). The TensorFlow version of the model was contributed by [sayakpaul](https://huggingface.co/sayakpaul). The original code and weights can be found [here](https://github.com/apple/ml-cvnets).
+
+
+## MobileViTConfig
+
+[[autodoc]] MobileViTConfig
+
+## MobileViTFeatureExtractor
+
+[[autodoc]] MobileViTFeatureExtractor
+    - __call__
+
+## MobileViTModel
+
+[[autodoc]] MobileViTModel
+    - forward
+
+## MobileViTForImageClassification
+
+[[autodoc]] MobileViTForImageClassification
+    - forward
+
+## MobileViTForSemanticSegmentation
+
+[[autodoc]] MobileViTForSemanticSegmentation
+    - forward
+
+## TFMobileViTModel
+
+[[autodoc]] TFMobileViTModel
+    - call 
+
+## TFMobileViTForImageClassification
+
+[[autodoc]] TFMobileViTForImageClassification
+    - call 
+
+## TFMobileViTForSemanticSegmentation
+
+[[autodoc]] TFMobileViTForSemanticSegmentation
+    - call 

docs/source/en/model_doc/mt5.mdx

@@ -96,3 +96,7 @@ See [`T5TokenizerFast`] for all details.
 ## FlaxMT5ForConditionalGeneration
 
 [[autodoc]] FlaxMT5ForConditionalGeneration
+
+## FlaxMT5EncoderModel
+
+[[autodoc]] FlaxMT5EncoderModel

docs/source/en/model_doc/mvp.mdx

@@ -0,0 +1,138 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# MVP
+
+## Overview
+
+The MVP model was proposed in [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
+
+
+According to the abstract,
+
+- MVP follows a standard Transformer encoder-decoder architecture.
+- MVP is supervised pre-trained using labeled datasets.
+- MVP also has task-specific soft prompts to stimulate the model's capacity in performing a certain task.
+- MVP is specially designed for natural language generation and can be adapted to a wide range of generation tasks, including but not limited to summarization, data-to-text generation, open-ended dialogue system, story generation, question answering, question generation, task-oriented dialogue system, commonsense generation, paraphrase generation, text style transfer, and text simplification. Our model can also be adapted to natural language understanding tasks such as sequence classification and (extractive) question answering.
+
+Tips:
+- We have released a series of models [here](https://huggingface.co/models?filter=mvp), including MVP, MVP with task-specific prompts, and multi-task pre-trained variants.
+- If you want to use a model without prompts (standard Transformer), you can load it through `MvpForConditionalGeneration.from_pretrained('RUCAIBox/mvp')`.
+- If you want to use a model with task-specific prompts, such as summarization, you can load it through `MvpForConditionalGeneration.from_pretrained('RUCAIBox/mvp-summarization')`.
+- Our model supports lightweight prompt tuning following [Prefix-tuning](https://arxiv.org/abs/2101.00190) with method `set_lightweight_tuning()`.
+
+This model was contributed by [Tianyi Tang](https://huggingface.co/StevenTang). The detailed information and instructions can be found [here](https://github.com/RUCAIBox/MVP).
+
+## Examples
+For summarization, it is an example to use MVP and MVP with summarization-specific prompts.
+
+```python
+>>> from transformers import MvpTokenizer, MvpForConditionalGeneration
+
+>>> tokenizer = MvpTokenizer.from_pretrained("RUCAIBox/mvp")
+>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp")
+>>> model_with_prompt = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp-summarization")
+
+>>> inputs = tokenizer(
+...     "Summarize: You may want to stick it to your boss and leave your job, but don't do it if these are your reasons.",
+...     return_tensors="pt",
+... )
+>>> generated_ids = model.generate(**inputs)
+>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+["Why You Shouldn't Quit Your Job"]
+
+>>> generated_ids = model_with_prompt.generate(**inputs)
+>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+["Don't do it if these are your reasons"]
+```
+
+For data-to-text generation, it is an example to use MVP and multi-task pre-trained variants.
+```python
+>>> from transformers import MvpTokenizerFast, MvpForConditionalGeneration
+
+>>> tokenizer = MvpTokenizerFast.from_pretrained("RUCAIBox/mvp")
+>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp")
+>>> model_with_mtl = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mtl-data-to-text")
+
+>>> inputs = tokenizer(
+...     "Describe the following data: Iron Man | instance of | Superhero [SEP] Stan Lee | creator | Iron Man",
+...     return_tensors="pt",
+... )
+>>> generated_ids = model.generate(**inputs)
+>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+['Stan Lee created the character of Iron Man, a fictional superhero appearing in American comic']
+
+>>> generated_ids = model_with_mtl.generate(**inputs)
+>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+['Iron Man is a fictional superhero appearing in American comic books published by Marvel Comics.']
+```
+
+For lightweight tuning, *i.e.*, fixing the model and only tuning prompts, you can load MVP with randomly initialized prompts or with task-specific prompts. Our code also supports Prefix-tuning with BART following the [original paper](https://arxiv.org/abs/2101.00190).
+
+```python
+>>> from transformers import MvpForConditionalGeneration
+
+>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mvp", use_prompt=True)
+>>> # the number of trainable parameters (full tuning)
+>>> sum(p.numel() for p in model.parameters() if p.requires_grad)
+468116832
+
+>>> # lightweight tuning with randomly initialized prompts
+>>> model.set_lightweight_tuning()
+>>> # the number of trainable parameters (lightweight tuning)
+>>> sum(p.numel() for p in model.parameters() if p.requires_grad)
+61823328
+
+>>> # lightweight tuning with task-specific prompts
+>>> model = MvpForConditionalGeneration.from_pretrained("RUCAIBox/mtl-data-to-text")
+>>> model.set_lightweight_tuning()
+>>> # original lightweight Prefix-tuning
+>>> model = MvpForConditionalGeneration.from_pretrained("facebook/bart-large", use_prompt=True)
+>>> model.set_lightweight_tuning()
+```
+
+## MvpConfig
+
+[[autodoc]] MvpConfig
+
+## MvpTokenizer
+
+[[autodoc]] MvpTokenizer
+
+## MvpTokenizerFast
+
+[[autodoc]] MvpTokenizerFast
+
+## MvpModel
+
+[[autodoc]] MvpModel
+    - forward
+
+## MvpForConditionalGeneration
+
+[[autodoc]] MvpForConditionalGeneration
+    - forward
+
+## MvpForSequenceClassification
+
+[[autodoc]] MvpForSequenceClassification
+    - forward
+
+## MvpForQuestionAnswering
+
+[[autodoc]] MvpForQuestionAnswering
+    - forward
+
+## MvpForCausalLM
+
+[[autodoc]] MvpForCausalLM
+    - forward

docs/source/en/model_doc/nezha.mdx

@@ -0,0 +1,76 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# Nezha
+
+## Overview
+
+The Nezha model was proposed in [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei et al.
+
+The abstract from the paper is the following:
+
+*The pre-trained language models have achieved great successes in various natural language understanding (NLU) tasks
+due to its capacity to capture the deep contextualized information in text by pre-training on large-scale corpora.
+In this technical report, we present our practice of pre-training language models named NEZHA (NEural contextualiZed
+representation for CHinese lAnguage understanding) on Chinese corpora and finetuning for the Chinese NLU tasks. 
+The current version of NEZHA is based on BERT with a collection of proven improvements, which include Functional 
+Relative Positional Encoding as an effective positional encoding scheme, Whole Word Masking strategy,
+Mixed Precision Training and the LAMB Optimizer in training the models. The experimental results show that NEZHA
+achieves the state-of-the-art performances when finetuned on several representative Chinese tasks, including
+named entity recognition (People's Daily NER), sentence matching (LCQMC), Chinese sentiment classification (ChnSenti)
+and natural language inference (XNLI).*
+
+This model was contributed by [sijunhe](https://huggingface.co/sijunhe). The original code can be found [here](https://github.com/huawei-noah/Pretrained-Language-Model/tree/master/NEZHA-PyTorch).
+
+## NezhaConfig
+
+[[autodoc]] NezhaConfig
+
+## NezhaModel
+
+[[autodoc]] NezhaModel
+    - forward
+
+## NezhaForPreTraining
+
+[[autodoc]] NezhaForPreTraining
+    - forward
+
+## NezhaForMaskedLM
+
+[[autodoc]] NezhaForMaskedLM
+    - forward
+
+## NezhaForNextSentencePrediction
+
+[[autodoc]] NezhaForNextSentencePrediction
+    - forward
+
+## NezhaForSequenceClassification
+
+[[autodoc]] NezhaForSequenceClassification
+    - forward
+
+## NezhaForMultipleChoice
+
+[[autodoc]] NezhaForMultipleChoice
+    - forward
+
+## NezhaForTokenClassification
+
+[[autodoc]] NezhaForTokenClassification
+    - forward
+
+## NezhaForQuestionAnswering
+
+[[autodoc]] NezhaForQuestionAnswering
+    - forward

docs/source/en/model_doc/nllb.mdx

@@ -0,0 +1,98 @@
+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# NLLB
+
+**DISCLAIMER:** If you see something strange, file a [Github Issue](https://github.com/huggingface/transformers/issues/new?assignees=&labels=bug&template=bug-report.yml) and assign
+@LysandreJik
+
+## Overview of NLLB
+
+The NLLB model was presented in [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by Marta R. Costa-jussà, James Cross, Onur Çelebi,
+Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula,
+Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John Hoffman, Semarley Jarrett, Kaushik Ram Sadagopan, Dirk Rowe, Shannon Spruit, Chau Tran, Pierre Andrews,
+Necip Fazil Ayan, Shruti Bhosale, Sergey Edunov, Angela Fan, Cynthia Gao, Vedanuj Goswami, Francisco Guzmán, Philipp Koehn, Alexandre Mourachko, Christophe Ropers,
+Safiyyah Saleem, Holger Schwenk, and Jeff Wang.
+
+The abstract of the paper is the following:
+
+*Driven by the goal of eradicating language barriers on a global scale, machine translation has solidified itself as a key focus of artificial intelligence research today.
+However, such efforts have coalesced around a small subset of languages, leaving behind the vast majority of mostly low-resource languages. What does it take to break the
+200 language barrier while ensuring safe, high quality results, all while keeping ethical considerations in mind? In No Language Left Behind, we took on this challenge by
+first contextualizing the need for low-resource language translation support through exploratory interviews with native speakers. Then, we created datasets and models aimed
+at narrowing the performance gap between low and high-resource languages. More specifically, we developed a conditional compute model based on Sparsely Gated Mixture of
+Experts that is trained on data obtained with novel and effective data mining techniques tailored for low-resource languages. We propose multiple architectural and training
+improvements to counteract overfitting while training on thousands of tasks. Critically, we evaluated the performance of over 40,000 different translation directions using
+a human-translated benchmark, Flores-200, and combined human evaluation with a novel toxicity benchmark covering all languages in Flores-200 to assess translation safety.
+Our model achieves an improvement of 44% BLEU relative to the previous state-of-the-art, laying important groundwork towards realizing a universal translation system.*
+
+This implementation contains the dense models available on release. Let us know via a GitHub issue if you would like to see the MoE models as well.
+
+This model was contributed by [Lysandre](https://huggingface.co/lysandre). The authors' code can be found [here](https://github.com/facebookresearch/fairseq/tree/nllb).
+
+## Generating with NLLB
+
+While generating the target text set the `forced_bos_token_id` to the target language id. The following
+example shows how to translate English to French using the *facebook/nllb-200-distilled-600M* model.
+
+Note that we're using the BCP-47 code for French `fra_Latn`. See [here](https://github.com/facebookresearch/flores/blob/main/flores200/README.md#languages-in-flores-200)
+for the list of all BCP-47 in the Flores 200 dataset.
+
+```python
+>>> from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
+
+>>> tokenizer = AutoTokenizer.from_pretrained("facebook/nllb-200-distilled-600M")
+>>> model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M")
+
+>>> article = "UN Chief says there is no military solution in Syria"
+>>> inputs = tokenizer(article, return_tensors="pt")
+
+>>> translated_tokens = model.generate(
+...     **inputs, forced_bos_token_id=tokenizer.lang_code_to_id["fra_Latn"], max_length=30
+... )
+>>> tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)[0]
+Le chef de l'ONU dit qu'il n'y a pas de solution militaire en Syrie
+```
+
+### Generating from any other language than English
+
+English (`eng_Latn`) is set as the default language from which to translate. In order to specify that you'd like to translate from a different language,
+you should specify the BCP-47 code in the `src_lang` keyword argument of the tokenizer initialization.
+
+See example below for a translation from romanian to german:
+
+```py
+>>> from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
+
+>>> tokenizer = AutoTokenizer.from_pretrained(
+...     "facebook/nllb-200-distilled-600M", use_auth_token=True, src_lang="ron_Latn"
+... )
+>>> model = AutoModelForSeq2SeqLM.from_pretrained("facebook/nllb-200-distilled-600M", use_auth_token=True)
+
+>>> article = "Şeful ONU spune că nu există o soluţie militară în Siria"
+>>> inputs = tokenizer(article, return_tensors="pt")
+
+>>> translated_tokens = model.generate(
+...     **inputs, forced_bos_token_id=tokenizer.lang_code_to_id["deu_Latn"], max_length=30
+... )
+>>> tokenizer.batch_decode(translated_tokens, skip_special_tokens=True)[0]
+UN-Chef sagt, es gibt keine militärische Lösung in Syrien
+```
+
+## NllbTokenizer
+
+[[autodoc]] NllbTokenizer
+    - build_inputs_with_special_tokens
+
+## NllbTokenizerFast
+
+[[autodoc]] NllbTokenizerFast

docs/source/en/model_doc/opt.mdx

@@ -39,9 +39,33 @@ The original code can be found [here](https://github.com/facebookresearch/metase
 [[autodoc]] OPTModel
     - forward
 
-
 ## OPTForCausalLM
 
 [[autodoc]] OPTForCausalLM
     - forward
 
+## TFOPTModel
+
+[[autodoc]] TFOPTModel
+    - call
+
+## TFOPTForCausalLM
+
+[[autodoc]] TFOPTForCausalLM
+    - call
+
+## OPTForSequenceClassification
+
+[[autodoc]] OPTForSequenceClassification
+    - forward
+
+## FlaxOPTModel
+
+[[autodoc]] FlaxOPTModel
+    - __call__
+
+
+## FlaxOPTForCausalLM
+
+[[autodoc]] FlaxOPTForCausalLM
+    - __call__

docs/source/en/model_doc/owlvit.mdx

@@ -0,0 +1,108 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# OWL-ViT
+
+## Overview
+
+The OWL-ViT (short for Vision Transformer for Open-World Localization) was proposed in [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby. OWL-ViT is an open-vocabulary object detection network trained on a variety of (image, text) pairs. It can be used to query an image with one or multiple text queries to search for and detect target objects described in text.
+
+The abstract from the paper is the following:
+
+*Combining simple architectures with large-scale pre-training has led to massive improvements in image classification. For object detection, pre-training and scaling approaches are less well established, especially in the long-tailed and open-vocabulary setting, where training data is relatively scarce. In this paper, we propose a strong recipe for transferring image-text models to open-vocabulary object detection. We use a standard Vision Transformer architecture with minimal modifications, contrastive image-text pre-training, and end-to-end detection fine-tuning. Our analysis of the scaling properties of this setup shows that increasing image-level pre-training and model size yield consistent improvements on the downstream detection task. We provide the adaptation strategies and regularizations needed to attain very strong performance on zero-shot text-conditioned and one-shot image-conditioned object detection. Code and models are available on GitHub.*
+
+## Usage
+
+OWL-ViT is a zero-shot text-conditioned object detection model. OWL-ViT uses [CLIP](clip) as its multi-modal backbone, with a ViT-like Transformer to get visual features and a causal language model to get the text features. To use CLIP for detection, OWL-ViT removes the final token pooling layer of the vision model and attaches a lightweight classification and box head to each transformer output token. Open-vocabulary classification is enabled by replacing the fixed classification layer weights with the class-name embeddings obtained from the text model. The authors first train CLIP from scratch and fine-tune it end-to-end with the classification and box heads on standard detection datasets using a bipartite matching loss. One or multiple text queries per image can be used to perform zero-shot text-conditioned object detection. 
+
+[`OwlViTFeatureExtractor`] can be used to resize (or rescale) and normalize images for the model and [`CLIPTokenizer`] is used to encode the text. [`OwlViTProcessor`] wraps [`OwlViTFeatureExtractor`] and [`CLIPTokenizer`] into a single instance to both encode the text and prepare the images. The following example shows how to perform object detection using [`OwlViTProcessor`] and [`OwlViTForObjectDetection`].
+
+
+```python
+>>> import requests
+>>> from PIL import Image
+>>> import torch
+
+>>> from transformers import OwlViTProcessor, OwlViTForObjectDetection
+
+>>> processor = OwlViTProcessor.from_pretrained("google/owlvit-base-patch32")
+>>> model = OwlViTForObjectDetection.from_pretrained("google/owlvit-base-patch32")
+
+>>> url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+>>> image = Image.open(requests.get(url, stream=True).raw)
+>>> texts = [["a photo of a cat", "a photo of a dog"]]
+>>> inputs = processor(text=texts, images=image, return_tensors="pt")
+>>> outputs = model(**inputs)
+
+>>> # Target image sizes (height, width) to rescale box predictions [batch_size, 2]
+>>> target_sizes = torch.Tensor([image.size[::-1]])
+>>> # Convert outputs (bounding boxes and class logits) to COCO API
+>>> results = processor.post_process(outputs=outputs, target_sizes=target_sizes)
+
+>>> i = 0  # Retrieve predictions for the first image for the corresponding text queries
+>>> text = texts[i]
+>>> boxes, scores, labels = results[i]["boxes"], results[i]["scores"], results[i]["labels"]
+
+>>> score_threshold = 0.1
+>>> for box, score, label in zip(boxes, scores, labels):
+...     box = [round(i, 2) for i in box.tolist()]
+...     if score >= score_threshold:
+...         print(f"Detected {text[label]} with confidence {round(score.item(), 3)} at location {box}")
+Detected a photo of a cat with confidence 0.707 at location [324.97, 20.44, 640.58, 373.29]
+Detected a photo of a cat with confidence 0.717 at location [1.46, 55.26, 315.55, 472.17]
+```
+
+This model was contributed by [adirik](https://huggingface.co/adirik). The original code can be found [here](https://github.com/google-research/scenic/tree/main/scenic/projects/owl_vit).
+
+## OwlViTConfig
+
+[[autodoc]] OwlViTConfig
+    - from_text_vision_configs
+
+## OwlViTTextConfig
+
+[[autodoc]] OwlViTTextConfig
+
+## OwlViTVisionConfig
+
+[[autodoc]] OwlViTVisionConfig
+
+## OwlViTFeatureExtractor
+
+[[autodoc]] OwlViTFeatureExtractor
+    - __call__
+
+## OwlViTProcessor
+
+[[autodoc]] OwlViTProcessor
+
+## OwlViTModel
+
+[[autodoc]] OwlViTModel
+    - forward
+    - get_text_features
+    - get_image_features
+
+## OwlViTTextModel
+
+[[autodoc]] OwlViTTextModel
+    - forward
+
+## OwlViTVisionModel
+
+[[autodoc]] OwlViTVisionModel
+    - forward
+
+## OwlViTForObjectDetection
+
+[[autodoc]] OwlViTForObjectDetection
+    - forward

docs/source/en/model_doc/pegasus_x.mdx

@@ -0,0 +1,45 @@
+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
+specific language governing permissions and limitations under the License.
+-->
+
+# PEGASUS-X
+
+## Overview
+
+The PEGASUS-X model was proposed in [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347)  by Jason Phang, Yao Zhao and Peter J. Liu.
+
+PEGASUS-X (PEGASUS eXtended) extends the PEGASUS models for long input summarization through additional long input pretraining and using staggered block-local attention with global tokens in the encoder.
+
+The abstract from the paper is the following:
+
+*While large pretrained Transformer models have proven highly capable at tackling natural language tasks, handling long sequence inputs continues to be a significant challenge. One such task is long input summarization, where inputs are longer than the maximum input context of most pretrained models. Through an extensive set of experiments, we investigate what model architectural changes and pretraining paradigms can most efficiently adapt a pretrained Transformer for long input summarization. We find that a staggered, block-local Transformer with global encoder tokens strikes a good balance of performance and efficiency, and that an additional pretraining phase on long sequences meaningfully improves downstream summarization performance. Based on our findings, we introduce PEGASUS-X, an extension of the PEGASUS model with additional long input pretraining to handle inputs of up to 16K tokens. PEGASUS-X achieves strong performance on long input summarization tasks comparable with much larger models while adding few additional parameters and not requiring model parallelism to train.*
+
+Tips:
+
+* PEGASUS-X uses the same tokenizer as PEGASUS.
+
+This model was contributed by [zphang](<https://huggingface.co/zphang). The original code can be found [here](https://github.com/google-research/pegasus).
+
+## PegasusXConfig
+
+[[autodoc]] PegasusXConfig
+
+
+## PegasusXModel
+
+[[autodoc]] PegasusXModel
+    - forward
+
+
+## PegasusXForConditionalGeneration
+
+[[autodoc]] PegasusXForConditionalGeneration
+    - forward

docs/source/en/model_doc/plbart.mdx

@@ -45,8 +45,9 @@ target text format is `[tgt_lang_code] X [eos]`. `bos` is never used.
 
 However, for fine-tuning, in some cases no language token is provided in cases where a single language is used. Please refer to [the paper](https://arxiv.org/abs/2103.06333) to learn more about this.
 
-In cases where the language code is needed, The regular [`~PLBartTokenizer.__call__`] will encode source text format, and it should be wrapped
-inside the context manager [`~PLBartTokenizer.as_target_tokenizer`] to encode target text format.
+In cases where the language code is needed, the regular [`~PLBartTokenizer.__call__`] will encode source text format 
+when you pass texts as the first argument or with the keyword argument `text`, and will encode target text format if
+it's passed with the `text_target` keyword argument.
 
 - Supervised training
 
@@ -56,11 +57,7 @@ inside the context manager [`~PLBartTokenizer.as_target_tokenizer`] to encode ta
 >>> tokenizer = PLBartTokenizer.from_pretrained("uclanlp/plbart-base", src_lang="en_XX", tgt_lang="python")
 >>> example_python_phrase = "def maximum(a,b,c):NEW_LINE_INDENTreturn max([a,b,c])"
 >>> expected_translation_english = "Returns the maximum value of a b c."
->>> inputs = tokenizer(example_python_phrase, return_tensors="pt")
->>> with tokenizer.as_target_tokenizer():
-...     labels = tokenizer(expected_translation_english, return_tensors="pt")
->>> inputs["labels"] = labels["input_ids"]
->>> # forward pass
+>>> inputs = tokenizer(example_python_phrase, text_target=expected_translation_english, return_tensors="pt")
 >>> model(**inputs)
 ```
 
@@ -88,7 +85,6 @@ inside the context manager [`~PLBartTokenizer.as_target_tokenizer`] to encode ta
 ## PLBartTokenizer
 
 [[autodoc]] PLBartTokenizer
-    - as_target_tokenizer
     - build_inputs_with_special_tokens
 
 ## PLBartModel

docs/source/en/model_doc/regnet.mdx

@@ -27,7 +27,8 @@ Tips:
 - One can use [`AutoFeatureExtractor`] to prepare images for the model.
 - The huge 10B model from [Self-supervised Pretraining of Visual Features in the Wild](https://arxiv.org/abs/2103.01988), trained on one billion Instagram images, is available on the [hub](https://huggingface.co/facebook/regnet-y-10b-seer)
 
-This model was contributed by [Francesco](https://huggingface.co/Francesco).
+This model was contributed by [Francesco](https://huggingface.co/Francesco). The TensorFlow version of the model
+was contributed by [sayakpaul](https://huggingface.com/sayakpaul) and [ariG23498](https://huggingface.com/ariG23498).
 The original code can be found [here](https://github.com/facebookresearch/pycls).
 
 
@@ -45,4 +46,15 @@ The original code can be found [here](https://github.com/facebookresearch/pycls)
 ## RegNetForImageClassification
 
 [[autodoc]] RegNetForImageClassification
-    - forward
+    - forward
+
+## TFRegNetModel
+
+[[autodoc]] TFRegNetModel
+    - call
+
+
+## TFRegNetForImageClassification
+
+[[autodoc]] TFRegNetForImageClassification
+    - call

docs/source/en/model_doc/resnet.mdx

@@ -31,7 +31,7 @@ The figure below illustrates the architecture of ResNet. Taken from the [origina
 
 <img width="600" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/resnet_architecture.png"/>
 
-This model was contributed by [Francesco](https://huggingface.co/Francesco). The original code can be found [here](https://github.com/KaimingHe/deep-residual-networks).
+This model was contributed by [Francesco](https://huggingface.co/Francesco). The TensorFlow version of this model was added by [amyeroberts](https://huggingface.co/amyeroberts). The original code can be found [here](https://github.com/KaimingHe/deep-residual-networks).
 
 ## ResNetConfig
 
@@ -47,4 +47,16 @@ This model was contributed by [Francesco](https://huggingface.co/Francesco). The
 ## ResNetForImageClassification
 
 [[autodoc]] ResNetForImageClassification
-    - forward
+    - forward
+
+
+## TFResNetModel
+
+[[autodoc]] TFResNetModel
+    - call
+
+
+## TFResNetForImageClassification
+
+[[autodoc]] TFResNetForImageClassification
+    - call

docs/source/en/model_doc/segformer.mdx

@@ -36,13 +36,14 @@ The figure below illustrates the architecture of SegFormer. Taken from the [orig
 
 <img width="600" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/segformer_architecture.png"/>
 
-This model was contributed by [nielsr](https://huggingface.co/nielsr). The original code can be found [here](https://github.com/NVlabs/SegFormer).
+This model was contributed by [nielsr](https://huggingface.co/nielsr). The TensorFlow version 
+of the model was contributed by [sayakpaul](https://huggingface.co/sayakpaul). The original code can be found [here](https://github.com/NVlabs/SegFormer).
 
 Tips:
 
-- SegFormer consists of a hierarchical Transformer encoder, and a lightweight all-MLP decode head.
+- SegFormer consists of a hierarchical Transformer encoder, and a lightweight all-MLP decoder head.
   [`SegformerModel`] is the hierarchical Transformer encoder (which in the paper is also referred to
-  as Mix Transformer or MiT). [`SegformerForSemanticSegmentation`] adds the all-MLP decode head on
+  as Mix Transformer or MiT). [`SegformerForSemanticSegmentation`] adds the all-MLP decoder head on
   top to perform semantic segmentation of images. In addition, there's
   [`SegformerForImageClassification`] which can be used to - you guessed it - classify images. The
   authors of SegFormer first pre-trained the Transformer encoder on ImageNet-1k to classify images. Next, they throw
@@ -51,6 +52,9 @@ Tips:
   found on the [hub](https://huggingface.co/models?other=segformer).
 - The quickest way to get started with SegFormer is by checking the [example notebooks](https://github.com/NielsRogge/Transformers-Tutorials/tree/master/SegFormer) (which showcase both inference and
   fine-tuning on custom data). One can also check out the [blog post](https://huggingface.co/blog/fine-tune-segformer) introducing SegFormer and illustrating how it can be fine-tuned on custom data.
+- TensorFlow users should refer to [this repository](https://github.com/deep-diver/segformer-tf-transformers) that shows off-the-shelf inference and fine-tuning.
+- One can also check out [this interactive demo on Hugging Face Spaces](https://huggingface.co/spaces/chansung/segformer-tf-transformers)
+  to try out a SegFormer model on custom images.
 - SegFormer works on any input size, as it pads the input to be divisible by `config.patch_sizes`. 
 - One can use [`SegformerFeatureExtractor`] to prepare images and corresponding segmentation maps
   for the model. Note that this feature extractor is fairly basic and does not include all data augmentations used in
@@ -65,7 +69,8 @@ Tips:
   used by [`SegformerForSemanticSegmentation`]). However, other datasets use the 0 index as
   background class and include this class as part of all labels. In that case, `reduce_labels` should be set to
   `False`, as loss should also be computed for the background class.
-- As most models, SegFormer comes in different sizes, the details of which can be found in the table below.
+- As most models, SegFormer comes in different sizes, the details of which can be found in the table below
+  (taken from Table 7 of the [original paper](https://arxiv.org/abs/2105.15203)).
 
 | **Model variant** | **Depths**    | **Hidden sizes**    | **Decoder hidden size** | **Params (M)** | **ImageNet-1k Top 1** |
 | :---------------: | ------------- | ------------------- | :---------------------: | :------------: | :-------------------: |
@@ -76,6 +81,10 @@ Tips:
 | MiT-b4            | [3, 8, 27, 3] | [64, 128, 320, 512] | 768                     | 62.6           | 83.6                  |
 | MiT-b5            | [3, 6, 40, 3] | [64, 128, 320, 512] | 768                     | 82.0           | 83.8                  |
 
+Note that MiT in the above table refers to the Mix Transformer encoder backbone introduced in SegFormer. For
+SegFormer's results on the segmentation datasets like ADE20k, refer to the [paper](https://arxiv.org/abs/2105.15203).
+
+
 ## SegformerConfig
 
 [[autodoc]] SegformerConfig
@@ -104,3 +113,23 @@ Tips:
 
 [[autodoc]] SegformerForSemanticSegmentation
     - forward
+
+## TFSegformerDecodeHead
+
+[[autodoc]] TFSegformerDecodeHead
+    - call
+
+## TFSegformerModel
+
+[[autodoc]] TFSegformerModel
+    - call 
+
+## TFSegformerForImageClassification
+
+[[autodoc]] TFSegformerForImageClassification
+    - call 
+
+## TFSegformerForSemanticSegmentation
+
+[[autodoc]] TFSegformerForSemanticSegmentation
+    - call 

docs/source/en/model_doc/speech-encoder-decoder.mdx

@@ -12,7 +12,7 @@ specific language governing permissions and limitations under the License.
 
 # Speech Encoder Decoder Models
 
-The [`SpeechEncoderDecoderModel`] can be used to initialize a speech-sequence-to-text-sequence model
+The [`SpeechEncoderDecoderModel`] can be used to initialize a speech-to-text model
 with any pretrained speech autoencoding model as the encoder (*e.g.* [Wav2Vec2](wav2vec2), [Hubert](hubert)) and any pretrained autoregressive model as the decoder.
 
 The effectiveness of initializing speech-sequence-to-text-sequence models with pretrained checkpoints for speech
@@ -20,9 +20,96 @@ recognition and speech translation has *e.g.* been shown in [Large-Scale Self- a
 Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli,
 Alexis Conneau.
 
-An example of how to use a [`SpeechEncoderDecoderModel`] for inference can be seen in
-[Speech2Text2](speech_to_text_2).
+An example of how to use a [`SpeechEncoderDecoderModel`] for inference can be seen in [Speech2Text2](speech_to_text_2).
 
+## Randomly initializing `SpeechEncoderDecoderModel` from model configurations.
+
+[`SpeechEncoderDecoderModel`] can be randomly initialized from an encoder and a decoder config. In the following example, we show how to do this using the default [`Wav2Vec2Model`] configuration for the encoder
+and the default [`BertForCausalLM`] configuration for the decoder.
+
+```python
+>>> from transformers import BertConfig, Wav2Vec2Config, SpeechEncoderDecoderConfig, SpeechEncoderDecoderModel
+
+>>> config_encoder = Wav2Vec2Config()
+>>> config_decoder = BertConfig()
+
+>>> config = SpeechEncoderDecoderConfig.from_encoder_decoder_configs(config_encoder, config_decoder)
+>>> model = SpeechEncoderDecoderModel(config=config)
+```
+
+## Initialising `SpeechEncoderDecoderModel` from a pretrained encoder and a pretrained decoder.
+
+[`SpeechEncoderDecoderModel`] can be initialized from a pretrained encoder checkpoint and a pretrained decoder checkpoint. Note that any pretrained Transformer-based speech model, *e.g.* [Wav2Vec2](wav2vec2), [Hubert](hubert) can serve as the encoder and both pretrained auto-encoding models, *e.g.* BERT, pretrained causal language models, *e.g.* GPT2, as well as the pretrained decoder part of sequence-to-sequence models, *e.g.* decoder of BART, can be used as the decoder.
+Depending on which architecture you choose as the decoder, the cross-attention layers might be randomly initialized.
+Initializing [`SpeechEncoderDecoderModel`] from a pretrained encoder and decoder checkpoint requires the model to be fine-tuned on a downstream task, as has been shown in [the *Warm-starting-encoder-decoder blog post*](https://huggingface.co/blog/warm-starting-encoder-decoder).
+To do so, the `SpeechEncoderDecoderModel` class provides a [`SpeechEncoderDecoderModel.from_encoder_decoder_pretrained`] method.
+
+```python
+>>> from transformers import SpeechEncoderDecoderModel
+
+>>> model = SpeechEncoderDecoderModel.from_encoder_decoder_pretrained(
+...     "facebook/hubert-large-ll60k", "bert-base-uncased"
+... )
+```
+
+## Loading an existing `SpeechEncoderDecoderModel` checkpoint and perform inference.
+
+To load fine-tuned checkpoints of the `SpeechEncoderDecoderModel` class, [`SpeechEncoderDecoderModel`] provides the `from_pretrained(...)` method just like any other model architecture in Transformers.
+
+To perform inference, one uses the [`generate`] method, which allows to autoregressively generate text. This method supports various forms of decoding, such as greedy, beam search and multinomial sampling.
+
+```python
+>>> from transformers import Wav2Vec2Processor, SpeechEncoderDecoderModel
+>>> from datasets import load_dataset
+>>> import torch
+
+>>> # load a fine-tuned speech translation model and corresponding processor
+>>> model = SpeechEncoderDecoderModel.from_pretrained("facebook/wav2vec2-xls-r-300m-en-to-15")
+>>> processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-xls-r-300m-en-to-15")
+
+>>> # let's perform inference on a piece of English speech (which we'll translate to German)
+>>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
+>>> input_values = processor(ds[0]["audio"]["array"], return_tensors="pt").input_values
+
+>>> # autoregressively generate transcription (uses greedy decoding by default)
+>>> generated_ids = model.generate(input_values)
+>>> generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+>>> print(generated_text)
+Mr. Quilter ist der Apostel der Mittelschicht und wir freuen uns, sein Evangelium willkommen heißen zu können.
+```
+
+## Training
+
+Once the model is created, it can be fine-tuned similar to BART, T5 or any other encoder-decoder model on a dataset of (speech, text) pairs.
+As you can see, only 2 inputs are required for the model in order to compute a loss: `input_values` (which are the
+speech inputs) and `labels` (which are the `input_ids` of the encoded target sequence).
+
+```python
+>>> from transformers import AutoTokenizer, AutoFeatureExtractor, SpeechEncoderDecoderModel
+>>> from datasets import load_dataset
+
+>>> encoder_id = "facebook/wav2vec2-base-960h"  # acoustic model encoder
+>>> decoder_id = "bert-base-uncased"  # text decoder
+
+>>> feature_extractor = AutoFeatureExtractor.from_pretrained(encoder_id)
+>>> tokenizer = AutoTokenizer.from_pretrained(decoder_id)
+>>> # Combine pre-trained encoder and pre-trained decoder to form a Seq2Seq model
+>>> model = SpeechEncoderDecoderModel.from_encoder_decoder_pretrained(encoder_id, decoder_id)
+
+>>> model.config.decoder_start_token_id = tokenizer.cls_token_id
+>>> model.config.pad_token_id = tokenizer.pad_token_id
+
+>>> # load an audio input and pre-process (normalise mean/std to 0/1)
+>>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
+>>> input_values = feature_extractor(ds[0]["audio"]["array"], return_tensors="pt").input_values
+
+>>> # load its corresponding transcription and tokenize to generate labels
+>>> labels = tokenizer(ds[0]["text"], return_tensors="pt").input_ids
+
+>>> # the forward function automatically creates the correct decoder_input_ids
+>>> loss = model(**input_features).loss
+>>> loss.backward()
+```
 
 ## SpeechEncoderDecoderConfig
 

docs/source/en/model_doc/speech_to_text.mdx

@@ -37,7 +37,7 @@ predicted token ids.
 
 The feature extractor depends on `torchaudio` and the tokenizer depends on `sentencepiece` so be sure to
 install those packages before running the examples. You could either install those as extra speech dependencies with
-`pip install transformers"[speech, sentencepiece]"` or install the packages seperately with `pip install torchaudio sentencepiece`. Also `torchaudio` requires the development version of the [libsndfile](http://www.mega-nerd.com/libsndfile/) package which can be installed via a system package manager. On Ubuntu it can
+`pip install transformers"[speech, sentencepiece]"` or install the packages separately with `pip install torchaudio sentencepiece`. Also `torchaudio` requires the development version of the [libsndfile](http://www.mega-nerd.com/libsndfile/) package which can be installed via a system package manager. On Ubuntu it can
 be installed as follows: `apt install libsndfile1-dev`
 
 
@@ -120,7 +120,6 @@ See the [model hub](https://huggingface.co/models?filter=speech_to_text) to look
     - save_pretrained
     - batch_decode
     - decode
-    - as_target_processor
 
 ## Speech2TextModel
 

docs/source/en/model_doc/speech_to_text_2.mdx

@@ -114,7 +114,6 @@ See [model hub](https://huggingface.co/models?filter=speech2text2) to look for S
     - save_pretrained
     - batch_decode
     - decode
-    - as_target_processor
 
 ## Speech2Text2ForCausalLM